That model remains relevant, but it is increasingly inadequate for the technologies attracting investment today. Artificial intelligence, robotics, biotechnology and advanced manufacturing depend on infrastructure, industrial capabilities and regulatory environments that cannot be replicated as easily as software. Their development is also becoming more exposed to geopolitical competition, as governments seek greater control over strategic technologies and supply chains.

The next generation of technology companies may be more technologically sophisticated than their predecessors, yet considerably harder to build.

When innovation meets industrial reality

A technological breakthrough does not necessarily translate into a commercially viable business. A robotics company may demonstrate impressive autonomous capabilities while struggling with manufacturing costs and industrial integration. A biotechnology startup may generate promising clinical results yet remain years away from regulatory approval and reimbursement. Even AI, often associated with software economics, increasingly depends on energy-intensive data centers, semiconductor supply chains and substantial infrastructure investment.

These constraints fundamentally change how technology companies must be evaluated. Scientific excellence may establish technological potential, but manufacturing readiness, regulatory sequencing and commercial adoption ultimately determine whether that potential becomes economic value.

The real challenge in frontier technology is increasingly the distance between scientific validation and commercial viability. A technology can work perfectly and still fail economically if the industrial, financial and regulatory conditions required for deployment are not in place.

Capital structure becomes a competitive advantage

Traditional venture capital was largely refined around businesses capable of rapid expansion, relatively low marginal costs and successive financing rounds tied to commercial growth. Frontier technology often follows a different trajectory. Development cycles extend over years, substantial capital may be required before meaningful revenues emerge, and progress depends on milestones that rarely align with conventional fundraising calendars.

More venture capital is not necessarily the answer. Industrial partnerships, public funding, equipment financing and milestone-based investment can complement equity capital, allowing companies to match financing instruments with different stages of technological and commercial risk. Government-backed programmes are also becoming more consequential as industrial policy increasingly directs capital towards strategically important technologies.

The ability to structure capital around the actual economics of technological development, while navigating an increasingly politicized investment environment, may therefore become as important as identifying promising innovations in the first place.

A new geography of technological power

The geography of innovation is being reshaped by two competing forces. Major economic powers, including the U.S., China and the European Union, are increasingly using industrial policy, regulation, subsidies and market access to strengthen their positions in strategic technologies. Semiconductor export controls, investment screening and technology sovereignty initiatives illustrate how technological development has become an instrument of economic and geopolitical competition.

At the same time, internationally connected hubs such as Singapore, Dubai and Hong Kong are developing a different proposition. Rather than competing primarily through the size of their domestic markets, they seek to attract capital, talent, research activities and regional headquarters by offering access to multiple economic ecosystems. Their value increasingly lies in facilitating international coordination at a time when larger markets are becoming more protective of strategic capabilities.

For deep tech companies, this creates a more complex geography of opportunity. A business may develop intellectual property in Europe, source components from Asia, raise capital through international financial centers and commercialize its technology in the United States. But these decisions can no longer be made on cost and efficiency alone. Regulatory compatibility, geopolitical exposure and access to strategic supply chains increasingly influence corporate structures and investment decisions.

The resulting competitive advantage may lie not in concentrating every capability within one jurisdiction, but in building an international architecture that combines specialized resources, access to major markets and resilience to geopolitical fragmentation.

A new definition of competitive advantage

Advances in AI and scientific tools are making certain forms of innovation faster and more accessible. Yet commercializing those innovations requires increasingly sophisticated combinations of capital, infrastructure, industrial partnerships and international coordination. This creates a paradox: technological innovation may become easier while building a successful technology company becomes harder.

For entrepreneurs and investors alike, scientific differentiation is only the beginning. The companies that succeed will be those capable of navigating industrial constraints, financial complexity and geopolitical uncertainty without compromising their capacity to innovate.

In the emerging deep tech economy, competitive advantage will increasingly depend not only on what companies invent, but on how effectively they organize the capital, industrial capabilities and international relationships required to bring those inventions to market.

Consider Talen Energy’s agreement to supply electricity to Amazon Web Services from its nuclear facility in Pennsylvania. The arrangement has been described as representing approximately $18 billion in contracted revenue. Yet public documentation available in September 2026 suggests that the utility’s planning covers only around three-quarters of the contracted volume, while compensation for electricity not drawn could be considerably lower than the headline value. The discrepancy is revealing. A contract can be legally enforceable without translating into the cash flows investors anticipate. And the evidence needed to identify that distinction may sit in an entirely different set of documents.

This is a familiar problem in frontier technology and infrastructure investing. Valuations frequently depend on assumptions about regulatory approvals, industrial capacity, commercial commitments or financing milestones that receive surprisingly little independent scrutiny.

AI is beginning to change that. Not by eliminating investment uncertainty, but by making it considerably cheaper to investigate.

The economics of verification are changing

Historically, sophisticated due diligence has been expensive. Large infrastructure funds can commission extensive technical investigations, while family offices and smaller investment firms often lack the resources to examine every material assumption behind a transaction. The resulting asymmetry is significant. A €5 million investment in an industrial technology company may depend on the same grid connections, regulatory approvals or manufacturing constraints as a much larger transaction, without benefiting from comparable scrutiny.

AI is narrowing this gap.

Regulatory filings, permit registers, subsidy decisions and technical documentation can now be examined and cross-referenced in minutes rather than days. Information that was theoretically public but practically inaccessible is becoming available to a much wider range of investors. Yet faster document processing does not automatically produce better investment decisions. A subsidy award is not necessarily a cash disbursement. A projected commissioning date may depend on several counterparties whose own commitments remain conditional. The challenge is to establish what the available evidence actually supports, rather than what a financial model assumes it means.

And this requires looking beyond individual documents.

The Risks Between the Documents

Consider an industrial deeptech company preparing its first commercial manufacturing facility. The tech has been validated, customer demand appears credible, public funding has been awarded and an industrial partner has been selected. Each element looks reassuring when examined separately.

But the financing may depend on a subsidy disbursement, which requires eligible expenditures, which depend on equipment procurement, which itself depends on the availability of capital. Meanwhile, customer deliveries assume that the facility will be commissioned on schedule.

A delay at one point can undermine the economics of the entire project. This is particularly relevant to frontier technologies, where commercial success depends on coordinating scientific development, industrial capabilities, regulatory processes and capital across multiple organizations, sometimes in different jurisdictions. Traditional diligence tends to examine these components through separate legal, financial and technical workstreams. What is less systematically investigated is whether their underlying assumptions are mutually consistent.

The distinction matters for valuation.

A projected commercialization date is not simply a milestone on a timeline. It determines financing requirements, expected revenues and potentially the dilution required to reach profitability. A six-month delay can therefore have consequences far beyond the operational inconvenience it initially suggests. The objective should be to identify the few dependencies that genuinely determine the investment outcome, establish who controls them and assess the consequences if they do not materialize as expected.

Three questions to ask before signing:

•⁠ ⁠Contracted is not drawn. A signed volume is only revenue when someone uses it. Who is committed to take it, where, and from when?

•⁠ ⁠Whose date is it? A date the company sets is a plan. A date a grid operator, regulator or supplier has committed to is a fact.

•⁠ ⁠Approved is not paid. Grants, subsidies and permits arrive in stages, with conditions. What has been paid, and what is still conditional?

That is a different exercise from producing an exhaustive inventory of risks. It is also one that AI can make considerably more accessible.

Towards a different model for due diligence

My experience as a deeptech investor, European Innovation Council jury member and entrepreneur coach at ESADE has repeatedly exposed the same blind spot: investors scrutinize the technology, market and management team, while treating critical execution assumptions as relatively settled. The problem is rarely a complete absence of information. More often, nobody has been specifically tasked with connecting the evidence to the economics of the transaction.

This is what led me to develop DeepRadar, an independent practice that tests the assumptions an energy, infrastructure or deeptech valuation depends on. Part of it is free: anyone can see what the public documents on a company or project actually commit to. Anyone can examine what public documents establish about a company or project, without commissioning a lengthy research exercise. The reasoning is straightforward: automated document analysis is becoming a commodity. Its long-term value is unlikely to reside in summarizing information that everyone can increasingly access.

The real opportunity is to make independent verification available earlier in the investment process, particularly for investors who cannot justify traditional consulting engagements. Such investigations need not be exhaustive. A focused review of a handful of material assumptions can inform valuation, financing conditions or transaction structure before commitments become irreversible.

Importantly, the exercise still requires human judgment. AI can identify inconsistencies and surface evidence, but interpreting contractual dependencies, assessing execution credibility and translating uncertainty into investment terms remain matters of expertise. Independence also matters. Founders, advisers and transaction participants have legitimate but sometimes conflicting incentives. An external assessment can challenge assumptions that might otherwise survive simply because everyone involved has an interest in moving forward.

Over time, I expect this approach to become a more routine part of investment decisions, particularly in transactions where technical complexity has historically made independent scrutiny prohibitively expensive. The implications extend beyond efficiency. As the cost of verification declines, investors will have fewer reasons to accept material assumptions without testing them, and companies will face greater expectations to substantiate the milestones supporting their valuations.

AI is not making due diligence obsolete. It is making superficial due diligence increasingly difficult to justify.

—

About the contributor

Eden Djanashvili is the founder of DeepRadar. She has served on the European Innovation Council jury, invested in deeptech companies and teaches and coaches at ESADE.

The recovery is now difficult to ignore. Hong Kong’s economy expanded by 5.1 per cent in the first half of 2026, its strongest half-year performance in almost five years. Merchandise exports increased by 42.5 per cent during the first eight months, supported by international demand for AI-related electronics. Capital markets have rebounded even more dramatically: by September, 118 companies had raised approximately HK$388 billion (US$50 billion) through initial public offerings, more than double the amount raised during the corresponding period in 2025. Technology companies, particularly those operating across the AI, semiconductor and advanced manufacturing value chains, have been central to this resurgence.

Yet this is not the restoration of the Hong Kong that existed before 2020. The city is re-emerging with a different relationship to China, a more interventionist state and a deliberate industrial ambition. In September 2026, the government unveiled its first Five-Year Plan for Economic and Social Development, aligning its priorities with China’s national planning framework. For a territory historically associated with market-led development, the shift is significant.

The interesting question is therefore not whether Hong Kong is back. It is what Hong Kong is becoming, and whether its new position within China’s industrial ecosystem can generate a different form of international competitive advantage.

From financial intermediary to industrial orchestrator

For much of its modern history, Hong Kong prospered by standing between economic systems. It connected Chinese manufacturers with international customers, mainland companies with foreign capital, and global investors with opportunities that were otherwise difficult to access. Its competitive advantage lay less in producing technology than in reducing the frictions associated with trade, finance and cross-border ownership.

That proposition has not disappeared, but its geography has changed. Hong Kong’s economic centre of gravity has moved decisively towards mainland China. The government’s new planning framework formalises this evolution, placing greater emphasis on integration with the Greater Bay Area and the development of technology-intensive industries. The Northern Metropolis project, along the Shenzhen border, is intended to bring research institutions, advanced production facilities and commercial infrastructure into a more integrated economic geography.

For frontier technology, this is more than an administrative realignment. A robotics company may require university research, specialised components, pilot customers and contract manufacturing. Semiconductor ventures depend on fabrication and packaging capabilities that few startups can reproduce. Biotechnology requires clinical infrastructure, regulatory expertise and specialised production facilities. These capabilities rarely coexist within a single jurisdiction.

Hong Kong cannot recreate them all, nor does it need to. Its opportunity lies in orchestrating capabilities distributed across the region: research and company formation in Hong Kong, engineering and manufacturing across Shenzhen and the wider Greater Bay Area, and access to international capital through Hong Kong’s financial markets. In this model, the border becomes part of the venture architecture rather than merely a constraint.

This also explains the government’s increasingly active role. Hong Kong has introduced three separate HK$10 billion programmes supporting academic research commercialisation, advanced manufacturing infrastructure and public investment intended to attract private capital into strategic industries. Rather than simply providing favourable market conditions, the government is attempting to coordinate the transition from scientific discovery to industrial production.

The economic rationale is understandable. Deep-tech ecosystems face coordination problems that private capital alone cannot always resolve. Research institutions, industrial customers, specialised infrastructure and long-duration financing must develop in parallel, even when no individual participant can capture the full benefits of investing first.

But coordination is not the same as competitiveness. Public funding can accelerate industrial development; it cannot substitute indefinitely for technological differentiation, manufacturing efficiency or commercial demand. Hong Kong’s challenge is to combine the advantages of industrial policy with the entrepreneurial experimentation and capital discipline that historically distinguished its economy.

The missing layer between science and industrial capital

Hong Kong has never lacked scientific capability. Its universities conduct internationally recognised research in AI, biotechnology, robotics, materials science and electronics. Its more persistent weakness has been converting that research into companies capable of reaching industrial and commercial scale.

This is a familiar problem in frontier technology. Scientific excellence creates technological possibilities, but not necessarily viable businesses. A discovery must become a reliable product, achieve acceptable manufacturing economics, navigate regulatory processes and secure adoption within existing industrial systems. Each transition introduces different risks, capital requirements and institutional dependencies.

Traditional venture capital is not always well suited to this progression. Fund lifecycles and follow-on financing assumptions were largely developed around businesses that could scale without factories, laboratories or lengthy qualification cycles. Industrial technologies often require a different combination of research funding, venture equity, strategic investment, project financing and access to public markets.

Hong Kong is beginning to assemble a more credible financing continuum. Chapter 18C of the Hong Kong Stock Exchange’s listing rules allows specialist technology companies, including qualifying pre-commercial businesses, to access public markets under conditions adapted to their development stage. In the first half of 2026, 13 companies raised approximately US$3.8 billion through the regime, bringing cumulative fundraising to nearly US$5 billion across 21 listings since its introduction in 2023. Another 14 companies had active applications by the end of June. (HKEX)

The emergence of listed AI developers such as MiniMax and Zhipu AI, alongside semiconductor and robotics companies, illustrates Hong Kong’s growing capacity to finance businesses whose economic value depends on substantial technological investment and uncertain paths to profitability.

This matters beyond individual IPOs. A credible public-market exit mechanism changes the economics of earlier-stage investment, potentially expanding the pool of capital available to companies before commercial maturity. Combined with government-backed programmes and the Greater Bay Area’s industrial infrastructure, Hong Kong could connect stages of technological development that remain fragmented in many other ecosystems.

Yet the distinction between financing technological ambition and financing industrial execution remains critical.

The AI investment cycle has accelerated Hong Kong’s recovery, but it may also obscure its limitations. Much of the city’s recent trade growth reflects electronic products moving through its commercial networks rather than technologies developed locally. Similarly, strong demand for AI-related equities does not automatically create the capabilities required to manufacture advanced sensors, industrial robotics or semiconductor equipment.

The difficult middle of the deep-tech journey remains the decisive test: financing companies after scientific feasibility has been established, but before manufacturing economics, regulatory approval or commercial adoption have been demonstrated.

Talent presents a related ambiguity. Hong Kong has attracted substantial numbers of professionals through targeted admission programmes, with mainland Chinese applicants representing the majority of successful candidates. This strengthens connections with Shenzhen’s entrepreneurial ecosystem, but international scientific collaboration and access to diverse talent remain essential to frontier innovation.

The question is whether Hong Kong can deepen its integration with China’s industrial capabilities without progressively narrowing the international networks that make those capabilities commercially valuable beyond China.

A new gateway, but not for everyone

For investors, Hong Kong’s transformation creates a proposition that is both more interesting and more complicated than a conventional economic rebound.

The opportunity lies in an unusual concentration of complementary assets. Hong Kong offers research universities, sophisticated financial markets, a convertible currency and a legal system familiar to international investors. Across the border sit Shenzhen’s technology companies, Dongguan’s manufacturing base and extensive networks of suppliers, engineers and industrial customers. Few regions combine scientific research, production capacity and public-market liquidity within such a concentrated geography.

For semiconductors, biotechnology, robotics, advanced manufacturing and energy technologies, this proximity can materially influence development costs, industrialisation timelines and capital requirements. Companies may be able to access manufacturing capabilities that would take years and substantial investment to reproduce elsewhere.

But industrial proximity and international accessibility are not equivalent advantages.

Hong Kong’s deeper integration with China makes it increasingly relevant to companies seeking access to Chinese technology, manufacturing and capital. At the same time, geopolitical fragmentation is changing the conditions under which those capabilities can be deployed internationally.

Export controls, restrictions on sensitive investments, competing technology standards and the growing separation of Chinese and Western technology ecosystems affect more than market access. They influence where intellectual property can be developed, which components can be sourced, who can finance a company and which customers it can ultimately serve.

This creates a fundamental distinction between companies that use Hong Kong as an international platform and those for which the city primarily provides a financial interface to an otherwise mainland-centred business.

For investors, the distinction is consequential. A company’s valuation cannot be separated from the industrial and regulatory architecture through which it intends to scale. Access to manufacturing capacity may improve capital efficiency while creating dependencies on restricted technologies, concentrated suppliers or jurisdictions that constrain future commercial opportunities.

Hong Kong’s competitive advantage will therefore depend not simply on the volume of capital it attracts, but on the quality, international accessibility and strategic resilience of the technological value chains it helps finance.

The economics of a more fragmented world

Hong Kong’s first era of prosperity was built on globalisation’s expanding connectivity. Its next may be shaped by the opposite phenomenon: a world in which technological capabilities, industrial infrastructure and capital are becoming more geographically concentrated and politically constrained. This does not necessarily diminish the economic value of intermediation. It changes its nature.

When markets are broadly accessible, intermediaries compete largely on efficiency, liquidity and transaction costs. When markets fragment, value increasingly accrues to institutions capable of navigating differences in regulation, financing, industrial standards and access to strategic capabilities.

Hong Kong is attempting to occupy precisely that position. Its integration with China gives it proximity to one of the world’s most sophisticated manufacturing ecosystems; its international financial infrastructure provides channels through which some of those capabilities can be financed and commercialised beyond the mainland.

But the two advantages are not automatically complementary. Deeper industrial integration may increase Hong Kong’s economic relevance within China while reducing its accessibility to certain international investors and technology markets. Conversely, preserving international credibility may require institutional differentiation that becomes more difficult as economic policy converges with Beijing’s priorities.

The real test is therefore not whether Hong Kong can recover its former status as a global gateway, but whether it can remain sufficiently connected to different systems to make their growing separation economically productive.

That is a more specialised role than the one Hong Kong historically occupied, and potentially a more valuable one for certain frontier technologies. It also implies a less universal investment proposition: companies will benefit differently depending on their technological dependencies, capital requirements, regulatory exposure and target markets.

Hong Kong’s second act will ultimately be judged less by IPO rankings or the scale of its industrial programmes than by the companies it enables to emerge—particularly those capable of converting regional scientific and manufacturing advantages into businesses with genuinely international reach.

The opportunity is not to recreate the frictionless globalisation on which the city once prospered. It is to develop a credible economic function in a world where those frictions have become structural.

Building globally is not simply about selling into international markets, opening subsidiaries or accessing foreign capital. Increasingly, it is about understanding where the capabilities required to industrialise a technology actually reside, and how to assemble them without surrendering strategic control. This distinction matters because the geography of innovation is diverging from the geography of production, while the geography of value capture follows yet another logic.

For much of the past two decades, the deep tech ecosystem has operated around a relatively intuitive proposition: scientific breakthroughs create technological advantages, which can subsequently be transformed into commercially valuable businesses through sufficient capital, engineering and execution. The model has produced remarkable companies, but it has also encouraged a somewhat misleading understanding of where competitive advantage originates. Venture investors have become sophisticated at assessing scientific differentiation, intellectual property, technology readiness and market potential. They have been considerably less systematic in evaluating the industrial and geographical conditions required to translate those technologies into economically viable products.

AI is making this imbalance more consequential. Computational biology, generative materials design, automated experimentation and increasingly powerful simulation tools are compressing parts of the discovery process. The number of potentially valuable technological solutions is expanding faster than the infrastructure required to validate, manufacture and commercialise them. Yet the physical world remains stubbornly resistant to the economics of software. Manufacturing yields cannot be improved through additional computing power alone. Clinical trials still require patients, protocols and regulatory oversight. Industrial equipment must be commissioned, supply chains qualified, and production processes stabilised.

The paradox is that accelerating technological discovery may actually increase the relative scarcity, and therefore the economic value, of industrial execution.

But there is another dimension to this transformation, one that receives considerably less attention. The capabilities required to build frontier technologies are not distributed uniformly across the world. Scientific expertise, industrial infrastructure, specialised talent, financing instruments and regulatory environments remain concentrated in distinct geographical ecosystems. Meanwhile, geopolitical fragmentation is making access to these capabilities increasingly conditional.

The result is a new competitive equation: companies must industrialise technologies faster, with increasingly complex capital requirements, while navigating a world in which the most efficient value chains are not necessarily the most politically acceptable. This creates a counterintuitive proposition: as the world becomes more fragmented, the most competitive frontier companies may need to become more internationally integrated, not less.

The industrialisation paradox

Deep tech has traditionally been associated with technological risk. Can the underlying science work? Can a lab discovery be translated into a functional prototype? Can the company establish sufficient intellectual property protection to defend its position? These questions remain essential, but they describe only part of the problem. A technology can be scientifically validated without being industrially reproducible. It can be industrially reproducible without being economically competitive. And it can be economically competitive without being commercially deployable in the jurisdictions where demand exists.

These distinctions are fundamental because each transition requires different capabilities, capital structures and organisational arrangements. Northvolt illustrates the consequences of underestimating them. The Swedish battery manufacturer attracted more than $15 billion in financing and became one of Europe’s most prominent industrial projects. Its ambition was not simply to develop competitive battery technology, but to establish a European manufacturing ecosystem capable of challenging established Asian suppliers. Yet building industrial capacity proved considerably more difficult than financing it. Production delays, manufacturing difficulties and insufficient output undermined customer relationships and financial sustainability, culminating in the bankruptcy of its Swedish parent in March 2025.

The broader lesson is not that Europe cannot manufacture batteries, nor that Northvolt’s technological foundations were necessarily inadequate. It is that industrial competence cannot be purchased simply by financing industrial infrastructure. Manufacturing is an accumulated capability, dependent on process knowledge, engineering experience, supplier coordination, quality management and thousands of operational adjustments that rarely appear in investor presentations.

China’s competitive advantage in batteries and electric vehicles is frequently described through the lenses of subsidies, labour costs or production capacity. These factors matter, but they do not fully explain the performance of companies such as BYD or CATL. Their advantages also reflect decades of cumulative industrial learning, dense supplier networks and close integration between product development and manufacturing operations. Economists (from Alfred Marshall to Enrico Moretti) describe part of this phenomenon through agglomeration economies : productivity advantages arising from the geographical concentration of complementary activities. Specialised suppliers, engineers, equipment manufacturers and customers operate in proximity, reducing coordination costs and accelerating the circulation of practical knowledge.

These advantages are difficult to replicate because they are systemic rather than confined to individual companies. Scientific knowledge can increasingly circulate internationally through publications, software, talent mobility and collaborative research. Industrial knowledge, by contrast, is often tacit, embedded in production processes and dependent on repeated interactions between specialised actors.

The geography of knowledge is becoming more fluid. The geography of industrial competence remains considerably more concentrated.

This creates an important asymmetry between scientific and industrial development. A frontier company might possess a technically validated solution while remaining several years away from achieving commercially acceptable yields, unit economics or production reliability. The capital required to bridge that distance can exceed the investment necessary to establish the underlying technology. And as AI continues to accelerate discovery, the number of companies confronting this second transition is likely to increase.

The implication for investors is substantial. Technology readiness and industrial readiness should be assessed as separate dimensions of company development. A successful laboratory demonstration establishes that a solution is technically possible; it does not establish that the company can manufacture it economically, reproduce its performance at scale or integrate it into an existing industrial value chain. In many frontier sectors, these subsequent transitions may ultimately determine more of the company’s value than the initial scientific breakthrough.

The new geography of competitive advantage

The industrialisation challenge becomes more complex when considered alongside the fragmentation of international economic relations. The global economy is not simply dividing into competing blocs. It is developing an increasingly intricate structure of dependencies, restrictions and strategic complementarities.

The US remains exceptionally strong in VC financing, advanced computing, frontier research and the commercialisation of high-value technologies. China combines large domestic markets with extraordinary manufacturing capabilities and increasingly sophisticated scientific research. Europe retains considerable strengths in industrial engineering, advanced equipment, pharmaceuticals and specialised manufacturing. Japan, South Korea, Taiwan and Singapore occupy critical positions across semiconductor production, precision engineering, electronics and advanced industrial infrastructure.

These are not interchangeable ecosystems. Each offers a different combination of capabilities, costs, financing conditions and institutional constraints. Consequently, the most attractive jurisdiction for scientific development may differ substantially from the most attractive jurisdiction for industrial validation, manufacturing or commercial deployment.

Semiconductors provide perhaps the clearest illustration. For decades, the industry’s development has depended on a highly specialised international division of labour. American companies dominate important segments of chip design and electronic design automation. TSMC has developed world-leading manufacturing capabilities in Taiwan. ASML, headquartered in the Netherlands, supplies essential lithography equipment. Japanese companies occupy critical positions in semiconductor materials and manufacturing equipment. The competitive strength of this system comes from specialisation and coordination rather than complete national self-sufficiency.

Yet its geographical concentration has also become a strategic vulnerability. TSMC’s expansion into Arizona illustrates the resulting trade-off. Its first Arizona fabrication plant began high-volume production using 4-nanometre technology in late 2024, while the company has committed to expanding its American manufacturing footprint. The investment responds to commercial demand, supply-chain resilience and geopolitical considerations. But reproducing advanced semiconductor manufacturing outside Taiwan involves more than constructing fabrication facilities. It requires developing supplier ecosystems, training specialised workforces, establishing operational routines and integrating production into an existing global value chain.

The relevant economic question is therefore not simply whether semiconductor production can be relocated. It is whether the additional resilience and strategic autonomy justify the capital expenditure, operational complexity and potentially higher production costs. Technological sovereignty and industrial efficiency are not always aligned. Governments may rationally accept higher costs to reduce strategic dependencies. Private companies, however, must reconcile these objectives with capital efficiency, competitive pricing and shareholder returns.

Biotechnology reveals a different configuration of the same phenomenon. American and European pharmaceutical companies have increasingly turned to Chinese-developed drug candidates through licensing and strategic partnerships. These transactions reflect China’s growing capabilities in drug discovery and clinical development, as well as the commercial advantages of accessing differentiated assets without reproducing every stage of development internally. This development is significant because China is no longer simply a manufacturing location or an alternative market for Western pharmaceutical companies. It has become an increasingly important source of clinical-stage innovation, supported by domestic research capabilities, clinical infrastructure and a growing biotechnology ecosystem. For pharmaceutical groups facing substantial patent expirations, licensing promising assets developed in China can offer an economically attractive alternative to developing every programme internally.

The transactions also demonstrate that geopolitical competition does not necessarily eliminate economic interdependence. Companies may face political pressure to reduce strategic dependencies on China while simultaneously recognising the commercial advantages of accessing Chinese innovation. The relevant decision is not binary. It involves assessing which activities create unacceptable dependencies, which can be diversified, and which international relationships remain essential to competitive performance.

This is where the geography of innovation becomes inseparable from the geography of value capture. A technology may originate in one country, undergo clinical development in another, receive regulatory approval in a third and generate most of its economic value in a fourth. The country where a technology is invented is not necessarily the country where the greatest economic value is ultimately captured. And the company that owns the original intellectual property does not necessarily control the most valuable elements of the resulting commercial ecosystem.

In frontier tech, comparative advantage is no longer simply about where an activity can be performed most efficiently. It is about where that activity creates the greatest strategic value within the entire development pathway.

The strategic question is therefore changing. It is no longer simply where a company should locate its activities, but how it should sequence and connect them to maximise learning, preserve optionality and retain control over the value it creates.

The next industrial leaders will be architects

The natural response to these challenges might appear to be vertical integration. If industrial capabilities are scarce, supply chains vulnerable and international relationships increasingly uncertain, companies should internalise as much of their value chain as possible. There is considerable logic to this argument. BYD’s integration across batteries, power electronics and vehicle manufacturing has allowed it to coordinate innovation across components and systems. TSMC’s manufacturing expertise creates advantages that cannot be reduced to individual fabrication technologies. In both cases, accumulated operational knowledge and coordination capabilities reinforce technological competitiveness.

But vertical integration also introduces significant costs. Owning industrial infrastructure increases capital intensity, reduces organisational flexibility and exposes companies to the financial consequences of underutilised capacity. For young frontier companies, these costs can become particularly problematic because technological, commercial and regulatory uncertainties remain substantial.

The appropriate question is therefore not whether a company should integrate vertically. It is which activities must be controlled to preserve competitive advantage, and which can be accessed more efficiently through external ecosystems.

BioNTech’s collaboration with Pfizer provides a useful counterpoint to the assumption that frontier companies must independently own the entire industrial value chain. BioNTech contributed its mRNA technology and specialised expertise, while Pfizer brought extensive capabilities in clinical development, regulatory affairs, manufacturing and global distribution. The collaboration allowed the companies to combine complementary assets rather than requiring BioNTech to reproduce the infrastructure of a multinational pharmaceutical group before commercialising its technology.

The partnership was not simply an outsourcing arrangement. It represented a deliberate combination of capabilities, capital and organisational infrastructure, allowing both companies to accelerate development and deployment. The lesson is not that outsourcing is inherently superior to vertical integration, but that strategic control does not necessarily require ownership of every productive asset.

This is fundamentally a problem of organisational and economic design. A company may identify an excellent manufacturing partner in another jurisdiction but discover that the contractual arrangement creates unacceptable intellectual property exposure. It may accelerate clinical development internationally while generating evidence that does not fully satisfy the requirements of its intended commercial market. Each decision may appear rational in isolation while producing an inefficient overall architecture.

The deeper issue is that frontier companies operate through several interconnected systems. Their technological architecture determines how knowledge, components and processes interact. Their industrial architecture determines how products are manufactured, validated and delivered. Their financial architecture determines how risks are allocated and which forms of capital can support successive development stages. Their geographical architecture determines where capabilities are accessed and how regulatory, commercial and political constraints affect execution.

In this context, capital efficiency becomes more than a function of operating discipline. It becomes a consequence of how intelligently the company has designed its international development architecture. This is not simply expansion. It is company building at the level of the global value chain.

The next generation of frontier technology companies will operate in an environment characterised by two apparently contradictory trends. Technological knowledge will become increasingly accessible, supported by AI, computational infrastructure and international research networks. Meanwhile, the physical, financial and institutional capabilities required to commercialise that knowledge will remain unevenly distributed and increasingly exposed to geopolitical constraints.

Scientific differentiation will remain essential, but it may become insufficient to establish a durable competitive advantage. Industrial capabilities, regulatory expertise, manufacturing relationships and the ability to coordinate geographically distributed operations will become increasingly valuable. The companies that succeed will not necessarily be those that internalise the greatest number of activities or establish operations in the largest number of countries. They will be those that understand where critical capabilities reside, which dependencies are acceptable, which assets must remain under their control and how to assemble an economically coherent value chain across jurisdictions.

The next industrial revolution will not be won by the countries that invent everything, nor by the companies that attempt to manufacture everything. It will be shaped by those capable of identifying what must be controlled, understanding what can be accessed elsewhere, and assembling the right capabilities faster than their competitors.

In a fragmenting world, the ability to integrate may become the ultimate competitive advantage.

I have been building a medtech venture in the UAE for two years without institutional capital. Not slowly. Literally without it.

During that period, we still managed to assemble a multidisciplinary team, put strong research infrastructure in place, conducted early studies, filed a provisional patent and moved a first device towards prototyping. We have attracted senior medical and scientific profiles, built hardware and data science capabilities, completed first-wave user research and started discussions around clinical research partnerships in the UAE and the UK.

None of this is an argument against venture capital, nor am I suggesting that money is not needed. Quite the opposite. Building without it has forced me to develop a more precise understanding of what venture capital actually contributes, and what founders who lack it must learn to supply in its place.

It has also made me question a distinction that is too rarely made in early-stage investing: the difference between capital that enables something to begin and capital that accelerates something already in motion.

When money isn’t available, something else has to be

When it comes to early-stage team building, the transaction appears fairly straightforward: offer equity to compensate for below-market cash or, in some cases, no cash at all. That is correct as far as it goes. But it understates the role of the underlying mission in determining whether equity is sufficient compensation for a particular person at a particular moment.

The people who joined Yuthera without compensation are not primarily motivated by equity. Every one of us is motivated by a condition that medicine has largely ignored for decades: chronic pelvic pain, affecting hundreds of millions of women, for which the current standard of care can range from inadequate to actively harmful. Several members of the team have experienced the condition themselves. Most of us have watched people we care about manage it and suffer silently. The equity is real, and it matters. But it is not the primary mechanism. At least not for us.

This is obviously not a scalable hiring strategy for every company or every category. I would be dishonest if I suggested that every conversation has produced talent aligned with that motivation. Mission cannot permanently substitute for salaries, and conviction does not pay people’s bills. But for companies working on problems where the human stakes are visible and the need for better solutions is personally felt by those qualified to contribute, mission can temporarily function as a form of capital. That matters for investors evaluating pre-revenue teams. A founder’s ability to assemble talented people before being able to pay market salaries does not validate the product or the market. But it does reveal something about persuasion, leadership and the capacity to mobilise resources that do not yet sit on the balance sheet.

The same applies to infrastructure. The conventional framing of university-startup collaboration positions the university as a source of IP and the startup as the go-to-market vehicle. The university generates science; the startup takes it to market. The relationship is largely transactional. What this framing misses is a more interesting dynamic, particularly in emerging innovation ecosystems. In the UAE, universities are also attempting to demonstrate that they are commercially and institutionally relevant in a region actively building its own knowledge economy. This can create a window in which a founder with a credible problem, a coherent plan and the willingness to show up can access research resources, engineering expertise, regulatory knowledge and institutional credibility that would be extremely expensive to replicate commercially.

I discovered this in a fairly literal way. I drove to the University of Sharjah and Khalifa University without appointments. I arrived, explained what we were building and asked for research partnerships. At the University of Sharjah, I sat outside the research institute director’s office for hours before getting the opportunity to make the case. Those conversations ultimately opened access to fabrication and testing infrastructure, biomedical engineering expertise, data architecture capabilities and government-linked prototyping resources. The infrastructure was already there. What was missing was a sufficiently credible reason to deploy it.

That distinction matters. Resource-constrained founders are constantly doing some version of this: identifying resources they cannot afford to purchase and finding another mechanism through which they can be mobilised. Sometimes that mechanism is equity. Sometimes it is mission. Sometimes it is institutional alignment, reputation, partnership or simple persistence.

Money is one mechanism for allocating resources. It is not always the only one.

The best capital accelerates what already works

Building without capital does not diminish the contribution capital makes at the right stage. If anything, it clarifies it. Where venture capital becomes most valuable is not necessarily as an enabling condition, but as an acceleration mechanism. The question is not simply whether a startup can exist without institutional funding. Some can, and for considerably longer than investors might assume. The more useful question is whether the company can reach its next threshold of technical, commercial or institutional credibility within the time available.

For a medical-device company such as Yuthera, that threshold might be a working prototype followed by the evidence required to move towards clinical validation. For a SaaS company, it might be the first hundred enterprise customers. Capital accelerates the journey towards those thresholds. But it does not necessarily determine whether the threshold is reachable. This creates an important distinction for both founders and investors. Founders who understand precisely what capital will accelerate tend to make a more compelling fundraising argument. They can explain what already exists, what the new capital unlocks, what milestones it should produce and why reaching those milestones faster is worth the dilution.

The alternative is a fundraising proposition in which almost everything begins after the cheque arrives: once funded, we will hire the team; once funded, we will build the prototype; once funded, we will discover whether the organisation can actually execute. From an investor’s perspective, that is a materially different risk. This does not mean every founder should bootstrap a deep-tech company for years. In many categories, that would be impossible. But early resourcefulness provides investors with information. It says something about what might be called a founder’s conversion efficiency: the amount of organisational and technical progress that can be produced from a limited pool of financial resources. That is particularly relevant in deep tech, where the capital requirements eventually become substantial. The most valuable venture capital is therefore not necessarily the largest cheque or even the earliest one. It is capital arriving when the founder has demonstrated enough resourcefulness to show that the money will accelerate something real rather than merely initiate something uncertain. That should change both how founders raise and how investors underwrite them.

Eventually, resourcefulness meets startup physics

There is, however, a danger in romanticising all of this. Resourcefulness is a great asset, but it is not an infinite substitute for capital. Founders who confuse resilience with strategy will eventually discover the difference, often at considerable cost.

Clinical trials require capital. Not because the science is unavailable or patients are unwilling, but because regulatory processes require documentation, oversight, quality systems and institutional infrastructure with genuine costs. Regulatory submissions require capital. International expansion requires capital. Manufacturing at the scale necessary to convert a working prototype into a commercially viable product almost always requires capital. And mission-driven hires still have mortgages, families and living expenses. Mission does not pay those bills indefinitely. There is therefore a point at which capital scarcity stops creating discipline and starts creating delay.

A founder who finds a creative way to avoid spending $100,000 while achieving essentially the same result may have demonstrated excellent capital efficiency. A founder who spends six additional months avoiding an expenditure that could have compressed the same milestone into six weeks may simply be running an undercapitalised company. That distinction is critical.

Capital scarcity is useful when it forces prioritisation. It becomes destructive when it forces delay.

The appropriate conclusion is therefore not that capital is unnecessary. It is that a founder’s demonstrated ability to build without it changes the terms on which capital should be sought, and, potentially, the terms on which it should be offered. A company that has assembled a credible team, established research partnerships, attracted clinical advisers and moved towards a working prototype without institutional funding is not asking an investor for its first injection of belief. It is asking for an accelerant for a fire that is already burning.

That is a materially different ask. For founders, it changes the fundraising conversation from “here is what we will begin once you fund us” to “here is what we have already demonstrated, here is the constraint we have now reached, and here is precisely what your capital allows us to accelerate.”

For investors, the question should change too. Not simply: What will you build if we fund you? But: What have you already learned how to build without us, and what becomes possible once time, rather than money, becomes the scarce resource?

This is not the usual account of executive productivity, which tends to focus on better prioritisation, delegation, inbox management and the disciplined use of increasingly elaborate task systems. These methods can be useful because leadership produces more obligations than memory can safely contain, but they address the administration of work rather than the temporal conditions under which difficult work becomes possible. A CEO may know precisely what matters, maintain an immaculate list and still reach Friday without having spent two continuous hours on the question most likely to determine the company’s trajectory.

When every meeting makes sense and the week does not

I am sure many of you know this very well. The calendar fills for legitimate reasons. Customers need attention, investors request updates, candidates must be interviewed, senior managers require decisions and problems that cannot be resolved elsewhere travel upwards. In frontier tech companies, these demands are intensified by the interdependence of technical, regulatory, industrial and financial risk. The CEO is often one of the few individuals capable of seeing the complete system, which makes their involvement rational and, beyond a certain point, operationally destructive.

Each meeting can make sense while the week they create does not. The ‘real’, deep work eventually migrates to evenings and weekends, not because these periods possess an inherent intellectual advantage, but because they offer something the formal week no longer provides: continuity.

The data suggests that this experience extends well beyond CEOs. Microsoft’s 2023 Work Trend Index, drawing on a survey of 31,000 people across 31 countries and activity across its productivity tools, found that 68 per cent of respondents lacked sufficient uninterrupted focus during the working day. Employees were spending 57 per cent of their time communicating through meetings, email and chat, compared with 43 per cent creating in documents, spreadsheets and presentations. In Singapore, the imbalance appeared more pronounced, with 82 per cent of employees reporting insufficient uninterrupted focus. Inefficient meetings ranked as the leading disruption to productivity, ahead even of the absolute number of meetings.

The figures describe a general workplace problem, but the cost is not evenly distributed. In frontier tech, the value of uninterrupted time rises with the density of the context required to make a decision. A scientist returning to an experimental question may need to recover the original hypothesis, the limitations of the method and the significance of several previous failures. A technical founder may have to hold architecture, manufacturability, customer requirements and regulatory consequences inside the same act of judgment. A CEO preparing for a financing decision must connect technical milestones, available runway, investor expectations and the operational cost of reaching the next point of proof. These are not collections of independent tasks that can be advanced interchangeably in twenty-minute intervals.

The time required to reconstruct this context is largely absent from conventional measures of productivity. The organisation sees an executive return to a document after a call and assumes that the work has resumed, although the first part of that return may consist entirely of reopening material, recovering assumptions and locating the precise question that the interruption displaced. Physical systems receive more understanding. Laboratory instruments require calibration, production lines have changeover times and cleanrooms impose entry procedures. Nobody interprets these constraints as evidence that the equipment lacks urgency. Cognitive setup costs are less visible and are therefore absorbed privately by the person performing the work.

Research on interrupted work has found that people frequently compensate by working faster, but with higher stress, frustration, effort and perceived time pressure rather than with an improvement in the underlying quality of the working environment. The organisation may therefore preserve output for a period while consuming more of the individual’s cognitive and emotional capacity to produce it. At executive level, this compensation often looks like commitment: documents completed late at night, decisions prepared over the weekend and unusual responsiveness maintained across every channel. The work gets done, allowing the operating architecture that displaced it to remain largely unexamined.

The CEO as the company’s favourite shortcut

No single meeting is responsible for this outcome, which is why it can persist even in well-managed companies. Each interruption is relatively small, each participant has a defensible reason for requesting attention and each discussion can plausibly be connected to an important objective. The cost becomes visible only in aggregate, through slower technical progress, prolonged decisions, incomplete reasoning and a growing dependence on work performed outside the hours the organisation has formally allocated.

The problem is not simply that the CEO has too many meetings. The calendar often reveals a deeper form of organisational dependence. When ownership is unclear, asking the CEO is safer than deciding. When functions struggle to reconcile competing objectives, inviting the CEO is easier than designing the interface between them. When information has not been documented, another meeting becomes the most convenient way to reconstruct it. The chief executive gradually becomes the company’s preferred solution to ambiguity: highly capable, immediately accessible and increasingly responsible for the latency of the entire system.

One of our portcos’ CEO has concentrated most internal reviews into a limited number of afternoon windows, requiring issues raised outside them to satisfy a clear threshold of urgency. Another reserves several mornings for product and technical work, while commercial and investor conversations are grouped later in the day. None has become inaccessible. The design simply distinguishes access from permanent interruptibility and prevents the organisation from using the CEO’s calendar as a substitute for clearer decision architecture.

The stronger versions of this practice extend beyond the chief executive. If only the CEO receives protected time while scientists, engineers and senior operators remain fragmented across the week, the arrangement is a personal privilege rather than an operating principle. The more interesting companies identify which forms of work require continuity and organise time accordingly. Laboratory teams preserve complete experimental sequences, technical leaders hold office hours rather than answering every question as it appears and written material precedes meetings in which several senior people are expected to make a decision.

The small delay introduced by these mechanisms is often productive. A question that must be formulated may answer itself. A problem that can wait for technical office hours may be resolved by the team before the meeting occurs. Several requests may reveal themselves as symptoms of the same weakness in documentation, training or decision authority. Scarce judgment remains available, but the company must use it with greater intention.

This does not require the elimination of meetings or the romantic isolation of technical work from the rest of the organisation. Frontier tech depends on coordination across functions whose evidence, incentives and time horizons differ considerably. The objective is not fewer meetings as an aesthetic preference, but a clearer distinction between communication that advances the work and communication that displaces it. A recurring meeting should make a decision, allocate resources, remove a constraint or review a prior commitment. If its principal function is to reassure participants that everybody remains informed, written communication may provide the information at a lower cost.

A calendar without slack is a hidden weekend plan

The CEOs who manage this well do not simply resolve to become more focused. They turn the preference into a system. Protected periods are visible in the calendar, understood by the organisation and enforced through the tools by which other people request time. Scheduling links do not offer unavailable hours, recurring meetings are concentrated within defined windows and exceptions are sufficiently unusual to remain recognisable as exceptions.

The technical enforcement matters because a boundary that must be defended separately against every reasonable request will eventually disappear. Scheduling software, unlike an accommodating executive, feels no discomfort when declining an appointment. This apparent rigidity tends to produce greater flexibility elsewhere. A CEO whose calendar is filled to capacity has no credible way to absorb the unexpected, even though the unexpected is one of the few predictable features of company building. If every hour has already been allocated, the issue can be managed only by delaying it, displacing existing work or extending the working day.

Protected time therefore serves two purposes. Under ordinary conditions, it provides continuity for work requiring analysis, writing or judgment. Under exceptional conditions, it becomes operating slack that can absorb a genuine emergency without transferring the cost to the evening or weekend.

A calendar without slack is therefore not a fully utilised system. It is a system whose contingency capacity has been quietly outsourced to nights and weekends.

Scarcity also improves the quality of coordination. When the CEO’s availability is treated as abundant, meetings become the default mechanism for transferring context, obtaining reassurance or avoiding local responsibility. When access is constrained, a subject that can wait becomes a message, a problem that belongs to a manager remains with the manager and a meeting without a clear object becomes more difficult to justify. The conversations that survive are more likely to involve decisions for which synchronous attention is genuinely useful.

Read the calendar before the strategy deck

Therefore, I would argue that the calendar is more than a personal productivity tool. It is an informal map of organisational dependency. It shows which decisions cannot travel without the CEO, which functions remain unable to integrate their work without escalation and which forms of progress the company is prepared to interrupt in order to maintain the appearance of coordination.

This becomes increasingly important as the organisation scales. In the early years, constant interaction can be efficient because context is shared, roles are fluid and the distance between a discussion and an action is small. As the company grows, the same habits become expensive. More people join each meeting, information must be reconstructed for participants who were not present earlier and senior leaders spend increasing amounts of time coordinating work they have decreasing amounts of time to examine.

For a board, the CEO’s calendar is therefore an operating document. It can reveal whether the organisation has distinguished urgency from importance, whether authority has genuinely been delegated and whether the company is preserving the cognitive capacity required for its hardest decisions. A chief executive who participates in every consequential conversation may appear deeply engaged, but engagement and leverage are not identical. The more mature operating model allows the CEO to remain available for the questions that require their judgment without becoming the surface on which every unresolved dependency is scheduled.

The best CEOs we encounter do not protect their mornings because they have discovered a superior productivity ritual. They do so because they understand that attention is an allocative decision. Every hour assigned to coordination is an hour unavailable for analysis, writing, product judgment or the integration of information that has not yet become a decision. The trade-off exists whether or not the organisation chooses to make it explicit.

A calendar cannot determine what a leader should think, but it determines whether they are likely to think about it for long enough. In companies attempting to turn uncertain science into reliable technology and reliable technology into an institution, that is not a question of personal efficiency.

It is part of the operating system.

Yet even at the height of the software era, entrepreneurial activity continued to gravitate towards a handful of ecosystems. Silicon Valley or the duo Boston-New York prospered because large markets, abundant capital, universities, exceptional talent and credible paths to liquidity reinforced one another in the same place.

Frontier tech startups are living into a rather different world. The chessboard has become more global, just as the world itself is becoming more balkanised.

A more global, but less frictionless, world

Scientific and engineering capabilities are now distributed everywhere. China and India have developed formidable pools of technical talent; Singapore has built a dense research environment; Gulf economies are investing heavily in tech and industrial capabilities; and Europe retains exceptional depth across science, engineering and advanced industry.

Frontier tech startups, however, depend on more than talent and capital. They need specialised infrastructure, manufacturing capacity, regulatory pathways, energy, supply chains and often access to governments or large industrial customers. Unlike software, these resources cannot simply follow entrepreneurs wherever they choose to congregate.

Besides, governments have now returned massively to the industrial arena: semicon, biotech, energy, and AI are increasingly treated not simply as markets, but as strategic capabilities. Subsidies, export controls, investment screening and domestic-production incentives can alter the relative attractiveness of locations while startups are still moving from laboratory validation towards industrial scale.

The paradox is striking: the resources required to build frontier companies have rarely been so geographically distributed, yet the freedom to combine them can no longer be taken for granted.

Europe has the ingredients, but scattered

Europe illustrates this tension particularly well. It combines leading universities, scientific talent, sophisticated industrial groups and deep engineering capabilities with comparatively fragmented capital markets and national innovation ecosystems. The conventional response has been to ask how Europe might reproduce the conditions that made Silicon Valley successful. A more useful question is what European companies should retain where the continent provides genuine differentiation, and what they should deliberately seek elsewhere.

Biotech offers a revealing example. In 2026, China’s Innovent and Pfizer agreed a partnership covering 12 oncology programmes in a transaction worth up to $10.5 billion. More interesting than its headline value is the division of labour: Innovent retains responsibility for development through Phase I, after which Pfizer assumes responsibility for global development. There is an economic logic behind this architecture. McKinsey estimates that Chinese biopharma companies can move from early discovery to IND 50 to 70 per cent faster than global benchmarks, while clinical trial recruitment can be two to five times faster than US and European benchmarks. Instead of insisting that an entire development chain remain within one ecosystem, the programme moves according to where each stage can be executed most effectively.

From international expansion to geographic architecture

This is more than internationalisation. Traditional international expansion begins with a successful company in a home market and asks where it should go next. Geographic architecture begins earlier, by asking where each critical component of the company ought to sit. Where is the scientific advantage? Where can validation be performed fastest? Which ecosystem offers competitive manufacturing? Which jurisdiction provides the most credible regulatory pathway? Where is patient capital available, and which beachhead market offers both early demand and strategic credibility? For frontier tech startups, these questions cannot be answered once in a business plan. The answers evolve as technology matures and as regulation, industrial policy and geopolitics reshape the relative advantages of different locations. The smartest European startups may therefore be those that remain European where their differentiation demands it, while being deliberately opportunistic about the architecture through which that advantage is industrialised.

Industrial policy is becoming part of product strategy

The same logic becomes more visible as companies approach production. Israeli battery tech company Addionics chose the United States for a planned $400 million manufacturing facility, not because the underlying technology originated there, but because customers, supply-chain policy and incentives under the Inflation Reduction Act had made American production increasingly attractive. For a software company, geography may affect hiring, taxation or market access. For an industrial company, it can change the economics of the product itself. Industrial policy is therefore becoming part of product strategy.

At the more geopolitical end of the spectrum, defence startup Covenant is developing production capacity across the United States, Germany and Israel. Its European industrial architecture is intended to limit dependence on American components, reflecting a world in which export controls, sovereign procurement requirements and political considerations can make a theoretically efficient global supply chain commercially or strategically unusable. The lesson is not that every activity should be distributed internationally. It is almost the opposite: every dependency now deserves greater scrutiny.

Investors need a different map too

This evolution changes the role of venture capital as well. Capital without an understanding of industrial geography is increasingly incomplete. Financing choices interact with intellectual property, manufacturing, regulation, governance, strategic partnerships and eventual market access, often across jurisdictions whose interests do not naturally align. This observation sits behind the cross-border investment thesis developed by Entropia Capital across Europe, APAC, the Gulf and North America. Rather than treating geography as a constraint surrounding an investment, the approach considers the configuration of ecosystems around a company as one of the variables through which value can be created.

The interesting opportunities often sit in the interstices: a European technology that can reach validation faster through an Asian partner; or an industrial process whose economics change once US incentives and customers are considered. Finding those combinations requires a different map from the one traditionally used by venture investors.

The Advantage Lies Between Ecosystems

The world is unlikely to return to the frictionless version of globalisation imagined at the beginning of the century. Strategic competition is intensifying, industrial policy has returned, and governments will remain active in determining where critical technologies are financed, manufactured and deployed. Nor does this imply that frontier innovation will become purely national. Few countries possess every scientific, industrial, financial and commercial capability required to dominate every emerging technology.

This creates the paradox that may define the next generation of frontier tech startups: they will need to become more sophisticated internationally precisely as the international system becomes more fragmented. For Europe, that need not be a counsel of pessimism. The continent possesses many of the pieces required to build globally significant companies, even if it does not always possess them in the same place, at the same scale or at the right moment.

The advantage may therefore belong to entrepreneurs capable of reading the global chessboard dynamically: knowing what must remain close to the scientific core, what can be sourced elsewhere, which partnerships create optionality, which dependencies create vulnerability and when geopolitical change requires the architecture to be redrawn.

The next great frontier companies won’t emerge from the ecosystem that has all the answers. They will emerge from those that become exceptionally good at connecting the right ones.

This article was first published on The European Business Review on 16 September 2026.

The disturbing paradox of intelligent failure

When John F. Kennedy entered the White House in 1961, he surrounded himself with what seemed to be an unbeatable concentration of intellectual talent. His advisers came from Harvard, Wall Street, the military and the commanding heights of American industry; they were articulate, analytical and supremely self-confident, and they shared their president’s conviction that intelligence, rationality and modern methods of management could solve problems that had defeated less sophisticated administrations.

Lyndon Johnson, Kennedy’s vice-president, was reportedly dazzled by them. Sam Rayburn, the veteran Speaker of the House and one of the most experienced political operators in Washington, was less reassured. After listening to Johnson describe the brilliance of the new administration, he observed that he would feel rather better if just one of these men had once run for county sheriff.

Rayburn was not defending ignorance against intelligence. He was pointing towards a distinction that is as important in business as it is in politics: the difference between intelligence as an individual faculty and judgement as a quality developed through contact with reality. The first can be demonstrated in an examination room, a laboratory or an investment committee; the second is acquired through exposure to consequences, conflicting interests, incomplete information and human beings who stubbornly refuse to behave as models predict.

In The Best and the Brightest, David Halberstam examined how Kennedy and Johnson’s exceptional advisers progressively involved the United States in Vietnam. The enduring interest of the book lies in the fact that its protagonists were neither fools nor obvious ideologues. They were serious, accomplished and, within their respective domains, highly competent men. Their failure was therefore considerably more unsettling than simple incompetence: they proved capable of deploying immense intelligence in defence of a mistaken representation of the world.

They interpreted Vietnam largely through the framework of the Cold War and the domino theory, reducing a conflict shaped by nationalism, colonial history, local legitimacy and Vietnamese political culture to another move in a global contest with communism. Once this model had been adopted, information was sorted according to whether it confirmed or disturbed it. Diplomats and observers with direct knowledge of the country were marginalised; quantitative indicators and military reports acquired an authority that their underlying assumptions did not merit; setbacks were interpreted not as evidence that the strategy might be mistaken, but as proof that it had not yet been applied with sufficient determination.

The result was a familiar escalation mechanism. Each additional commitment made the previous one appear more important, while the political cost of admitting error rose with the resources already invested. A policy that had never been chosen in a single, fully considered decision gradually became almost impossible to reverse. The administration repeatedly sought to “keep its options open”, yet every attempt to postpone a fundamental choice reduced the range of good options actually available.

Frontier tech companies should recognise themselves in this story, because they, too, deliberately assemble unusual concentrations of intelligence around difficult and uncertain problems. Their founders may be eminent professors, successful entrepreneurs or former executives from major industrial groups. Their boards include accomplished investors, scientists, public officials and corporate leaders. Their presentations contain patents, simulations, technical milestones, market forecasts and elaborate risk matrices, all of which convey an impression of disciplined mastery.

Nevertheless, many of these companies make decisions that appear, in retrospect, astonishingly detached from the evidence available inside their own organisations. They continue financing programmes after the scientific premise has weakened; they announce industrial timetables that their manufacturing teams regard as implausible; they treat regulatory objections as communication problems, customer resistance as a failure of education, and recurring technical anomalies as isolated incidents. Because the people involved are highly intelligent, they are rarely short of convincing explanations for doing so.

This is the paradox at the centre of deep-tech leadership: intelligence is indispensable, but it does not necessarily protect an organisation from error. Under certain conditions, it makes error more durable, because intelligent people are exceptionally capable of rationalising commitments, defending elegant models and producing sophisticated accounts of why contradictory evidence should not yet be believed.

Why frontier tech magnifies the problem

The expression frontier tech or “deep tech” covers very different sectors—biotechnology, semiconductors, robotics, new materials, quantum computing, aerospace, energy and advanced industrial systems—but these companies share a distinctive strategic structure. Their success depends not on one uncertainty but on a chain of interdependent uncertainties, each governed by a different body of knowledge and each capable of invalidating the economic proposition.

A scientific effect must first exist and be reproducible. It must then be transformed into an engineered system, manufactured at an acceptable yield, certified where necessary, incorporated into a customer’s operations, maintained in the field and sold at a price that supports the capital structure of the business. A company may be right about the science and wrong about industrialisation; right about the product and wrong about adoption; right about customer interest and wrong about the regulatory pathway; or right about all of these and still discover that the financing required to cross the gap between prototype and scale exceeds what its investors can provide.

No individual, regardless of ability, can master this entire chain. A distinguished physicist may understand a phenomenon better than anyone else in the world while possessing limited knowledge of quality systems, production economics or organisational design. A former executive from a global corporation may be highly effective at optimising mature operations but poorly adapted to a company in which even the fundamental technical architecture remains uncertain. Expertise, in other words, is local, whereas the CEO’s decisions are systemic. This creates an unavoidable information asymmetry: the person with the greatest authority rarely possesses the most relevant knowledge about every decision over which that authority is exercised.

In a healthy organisation, these partial views are combined into a progressively more accurate representation of reality. In an unhealthy one, they are ranked according to status: the abstraction held at the top defeats the observation made at the edge, even when the latter contains the information on which the company’s survival depends.

This is why the familiar distinction between vision and execution is particularly dangerous in deep tech. According to this managerial mythology, the leader defines the destination while the organisation handles implementation, which is often treated as a form of corporate plumbing: necessary, occasionally difficult, but fundamentally subordinate work that can be delegated once the architecture has been decided.

Yet in deep tech, the “plumbing” is precisely where the strategy encounters physics, regulation, economics and human behaviour. Whether a material can be produced consistently outside the laboratory, whether a sensor remains reliable under vibration and temperature variation, whether a biological process behaves similarly at ten thousand litres as it did at ten, whether a supplier can meet the required tolerances, or whether a customer can integrate the product without redesigning its operations are not details of execution. They determine whether the company possesses a business at all.

When technical accidents become organisational biographies

Catastrophic technological failures are usually described initially through their physical causes: a damaged heat shield, an erroneous software deployment. These causes matter, but investigations repeatedly reveal that the physical defect was only the final link in a longer organisational chain. The accident becomes, in effect, an involuntary biography of the institution that produced it.

Before the Challenger disaster in January 1986, engineers at Morton Thiokol expressed serious concern about the performance of the shuttle’s solid-rocket-booster seals in unusually cold weather. During the discussion preceding launch, Thiokol management reversed its initial recommendation and supported proceeding, contrary to the position of its engineers. The Rogers Commission subsequently identified failures of communication, a conflict between engineering evidence and managerial judgement, and a NASA structure that allowed safety issues to bypass senior shuttle managers. Challenger disintegrated 73 seconds after liftoff.

The case is often simplified into a morality tale in which engineers knew the truth and managers ignored them. The reality is more instructive. The available data were incomplete, the relationship between temperature and O-ring performance was contested, and a decision had to be made under operational pressure. This ambiguity did not reduce the need for caution; it made the quality of the decision-making process decisive. The failure lay partly in the way uncertainty was framed: instead of requiring evidence that launch was safe under unprecedented conditions, the discussion placed pressure on engineers to establish conclusively that it was unsafe.

Seventeen years later, the Columbia Accident Investigation Board concluded that the management practices overseeing the shuttle programme were as much a cause of the loss of Columbia as the foam strike that damaged its wing. Engineers had attempted to obtain better imagery of the orbiter in flight, but their concern did not produce an adequate response from the decision-making hierarchy. Schedule pressure, normalised anomalies and a fragmented safety structure again influenced the interpretation of ambiguous evidence. The board recommended an independent technical authority responsible for safety requirements and waivers, explicitly separated from responsibility for cost and schedule.

The repetition is more disturbing than either event considered separately. NASA was not an organisation that had never thought about safety; after Challenger, it had studied its failures extensively, revised procedures and pledged to learn. Columbia demonstrated that institutional learning is not permanent. Lessons decay as personnel change, commercial or political pressure returns, and practices introduced in response to a crisis become rituals whose original purpose is gradually forgotten.

An anomaly that does not produce a catastrophe can even weaken vigilance. Each successful mission in which foam was shed, or each launch in which an O-ring suffered damage without causing loss of vehicle, made the deviation appear more acceptable. The absence of disaster was misread as evidence of safety, although it might equally have been evidence of good fortune. This process, often described as the normalisation of deviance, is particularly dangerous in deep tech because systems can operate successfully for long periods while approaching a boundary that the organisation does not fully understand.

Theranos represents a more extreme and ethically different case, because its leadership crossed the boundary from excessive optimism into deception. Nevertheless, it demonstrates how prestige, secrecy and narrative control can isolate a technical company from corrective evidence. According to the SEC, Theranos raised more than $700 million while making false or exaggerated claims about its technology, business and financial performance. The company represented that its proprietary analyser could perform an extensive range of tests from very small blood samples; the SEC complaint stated that the device was used for only 12 of the tests offered to patients, with the majority performed on modified commercial equipment.

Theranos should not be used to imply that every technically overambitious founder is fraudulent. The more useful question is why an organisation operating in a field as demanding as medical diagnostics was governed in a way that allowed charisma, confidentiality and executive authority to become substitutes for independent technical verification. A prestigious board cannot compensate for insufficient domain knowledge, and loyalty to a mission cannot justify preventing qualified people from examining whether the mission’s central claims are true.

More recently, an expert panel examining Boeing’s safety culture reviewed thousands of pages of documents and interviewed more than 250 employees, managers, executives, supplier personnel and regulators. Among its observations was a particularly revealing criticism of Boeing’s “Seek, Speak & Listen” framework: the panel found considerable emphasis on speaking, but much less on seeking and listening. It also reported that the documentary and interview evidence did not provide objective support for a commitment to safety as foundational as the company’s descriptions suggested.

The distinction deserves attention. Many organisations now encourage employees to speak up, but the existence of a reporting channel says little about whether inconvenient information will influence a decision. Voice without attention is theatre; attention without authority is frustration; and authority without protection leaves the person who raises the concern dependent on the goodwill of precisely those whose plan is being challenged.

The real architecture of dissent

The usual response to these failures is to demand a better culture. Leaders announce that their doors are open, introduce whistleblowing systems, add “challenge” to the company’s values and remind employees that safety or integrity comes first. Such measures may be useful, but they often misunderstand the economics of speaking up.

For an employee, dissent is rarely costless. The benefit, i.e., preventing a future problem, is shared by the organisation and may never be visible, particularly if the warning succeeds. The potential cost (being labelled negative, obstructive, disloyal or insufficiently entrepreneurial) is immediate and personal. When the employee challenging the plan is junior, and the executive defending it controls compensation, promotion or continued employment, silence may be a perfectly rational response.

Leaders consequently learn much less from the absence of dissent than they imagine. A quiet meeting may indicate agreement, but it may equally indicate resignation, fear, fatigue or the belief that the decision has already been made. The more powerful and intellectually formidable the CEO, the more misleading apparent consensus can become, because disagreement carries both hierarchical and cognitive risk.

A functional architecture of dissent must therefore alter both process and power. For company-critical technical decisions, the person responsible for delivering the milestone should not be the sole judge of whether the evidence supporting it is adequate. Safety, technical integrity and regulatory compliance require authorities that possess sufficient independence to delay or stop a programme without being punished for the resulting effect on schedule.

The order of discussion also matters. If the CEO, scientific founder or most prestigious professor announces a view at the beginning of a meeting, everyone who follows must decide not only what they think, but whether contradicting that person is worth the social and professional cost. Asking the most junior or operationally exposed participants to speak first is not a matter of etiquette; it is a way of preventing hierarchy from contaminating the evidence before it has been collected.

Equally important is the treatment of weak signals. Employees should not be required to diagnose a complete failure mechanism before reporting an anomaly. In complex systems, the individual who notices that a measurement, sound, vibration, customer behaviour or software response is unusual may not be able to explain its significance. If the organisation demands a fully developed case before paying attention, it will systematically suppress the earliest and cheapest warnings.

The CEO should therefore be concerned with what might be called bad-news latency: the time separating the first observation of a serious problem from its arrival, in recognisable form, at the level where resources and priorities can be changed. In many failed organisations, the crucial information was not absent. It was delayed, diluted, reclassified or trapped within a part of the hierarchy that lacked the authority to act.

A doctrine of grounded leadership

The practical conclusion is not that deep tech CEOs should abandon vision, interfere in every experiment or replace expertise with instinct. Nor should organisations allow every objection to paralyse decisions, because uncertainty can never be eliminated and difficult technologies require persistence long after a conventional company would have withdrawn.

The distinction that matters is not between boldness and caution, but between conviction and dogmatism. Conviction holds that a problem is important enough to justify sustained effort and adaptation; dogmatism holds that the present explanation, architecture or timetable must be correct. Conviction permits learning because it attaches identity to the mission. Dogmatism prevents it because it attaches identity to the current plan.

The CEO must separate technical truth from organisational power. When a decision involves safety, regulatory compliance or a risk capable of threatening the company, an independent technical authority should be able to require more evidence or stop the process. That authority cannot remain meaningfully independent if career advancement depends entirely on the executive accountable for cost and schedule. NASA’s post-Columbia recommendation that technical authority be institutionally separated from programme delivery applies just as readily to a fusion company, a medical-device start-up or an autonomous-systems developer.

Capital allocation should be governed by the same discipline. When a programme misses successive milestones, management commonly responds by adding resources and urgency, partly because withdrawal would require admitting that earlier projections were wrong. Before increasing commitment, the board should require a precise account of what has been learned since the previous decision, which initial assumption has been invalidated, and why additional capital will change the underlying technical or commercial mechanism rather than merely finance another iteration of the same attempt.

This is particularly important because deep-tech progress is rarely linear. A missed milestone does not necessarily mean that a programme should be abandoned, just as a successful test does not necessarily mean that the uncertainty has been resolved. What matters is whether the company is producing information that changes the probability of success. A programme that fails intelligently may be more valuable than one that repeatedly achieves carefully selected milestones without addressing its principal risk.

Boards, for their part, must insist on proximity to evidence rather than reliance on narrative. They should periodically hear from technical, manufacturing, regulatory and commercial leaders without requiring every message to be translated by the CEO. They should examine distributions, reproducibility, yield, failure modes and confidence intervals rather than only averages and headline milestones. They should ask what evidence would disprove the investment thesis and whether the organisation is actively trying to obtain it.

The lesson of Halberstam’s “best and brightest” is therefore not that experts should be distrusted, that operational experience is always superior to analysis, or that complicated decisions should be surrendered to intuition. Deep tech companies could not exist without exceptional expertise, formal models and leaders capable of sustaining belief in achievements that initially appear improbable.

The lesson is that intelligence does not contain its own corrective mechanism. Brilliant people can construct more sophisticated reasons for ignoring contradictory evidence; homogeneous groups of experts can reinforce one another’s assumptions; detailed models can turn uncertainty into an illusion of precision; and an inspiring vision can become a means of avoiding an inconvenient present.

The decisive quality of deep-tech leadership is consequently not omniscience, but epistemic discipline: the ability to distinguish what the company knows from what it hopes, to recognise where one form of expertise ends and another begins, to remain close enough to operations that weak signals are still visible, and to revise a cherished belief without experiencing that revision as a personal defeat.

The company does not require a leader who is always right. It requires a leader who has made it institutionally difficult for everyone to remain wrong.

That is why due diligence cannot end with the question of whether management’s plan is credible. It rather challenges whether the world around the company will continue to give each milestone the meaning the investment case assumes. When markets are relatively stable, investors can compare a roadmap with familiar precedents: the usual duration of a qualification cycle, the recognised force of a contract, the normal availability of an industrial input. In a market being reshaped by scarcity, regulation or concentrated demand, those precedents can become unreliable before the language of investment committees has caught up.

The overheating of the data centre market makes this problem unusually visible. It is not the only sector in which it appears, and it is not the destination of the argument. It is a useful lens through which to examine a broader deep tech question we often face: what happens when the commercial promise remains legible, but the standards that once translated that promise into an investable asset begin to move?

When scarcity changes the milestone

A heated market does more than lift prices and lengthen queues. At a certain point, scarcity begins to alter the asset being bought. A grid connection may still be called a connection while offering a weaker claim on firm power. A supply agreement may secure a place in a manufacturer’s order book without securing the delivery date embedded in the revenue plan. A regulatory designation may remain valid while covering a narrower use, geography or production scale than the model assumes. The familiar milestone survives, but the economic right beneath it becomes conditional, divisible or exposed to a new counterparty.

This is a more difficult form of market risk than ordinary volatility. An increase in price is visible and can be reflected in a model. A change in meaning is easier to miss because the vocabulary remains intact. The company can accurately say that power is secured, a product is certified or capacity has been reserved, while the investor hears the stronger version of those claims inherited from an earlier market. The resulting distortion is rarely a simple falsehood. It is a gap between the promise contained in the roadmap and the rights that now sit behind the words used to describe it.

The data centre market makes the shift visible

PJM offers a clear example. In August 2026, the grid operator proposed allowing large new customers to connect even when the system had not secured enough generation to cover their full demand. The trade-off is simple: when power is scarce, the uncovered portion of their demand would be curtailed first. The proposal does not create more electricity. It changes what a grid connection means.

That distinction matters because PJM is already short of capacity. Its latest auction secured 138,318 MW and, even after counting supply committed outside the auction, left the system 6,831 MW below its own reliability requirement. Prices cleared at the regulatory ceiling. A project can therefore meet a milestone such as connected by Q3 2028 while still lacking an unconditional right to draw its full power requirement. The date has not moved. What the date represents has.

Recent events show how quickly these distinctions become material. In northern Virginia, a transmission fault in July triggered roughly 3,800 MW of data-centre demand to disconnect and switch to on-site generation. PJM concluded that the protective equipment had been set too cautiously: systems designed to protect individual facilities had become, at scale, a reliability issue for the wider grid. The equipment had not changed. The standard against which it was judged had.

Texas provides a different version of the same problem. Big Digital Energy announced a site with 311 MW of headline capacity. Its own disclosures reveal three very different assets beneath that number: 17 MW operating today; up to 111 MW of grid power still awaiting validation; and the balance dependent on gas generation that has yet to be built. Two weeks later, the company described the site as supporting up to 300 MW of total buildout. None of this makes the project weak, nor does it require assuming that anyone is misleading investors. The point is that a single headline can collapse three different states into one number: operating capacity, conditional capacity and future optionality. For an investor, they should not carry the same weight.

Diligence between the roadmap and the market

For early-stage investors, the implication is not to replace conviction with suspicion. It is to locate conviction more precisely. A deeptech roadmap already asks the investment team to judge technical progress, commercial adoption and the company’s ability to finance the journey between them.

The additional task is to identify which milestones depend on external systems whose own rules are changing. A certification date depends on the capacity and interpretation of a regulator. A scale-up plan depends on equipment lead times, specialist labour and suppliers that may be serving an entire wave of competing projects. A customer commitment depends on the legal force of the document, but also on the customer’s ability to absorb the product when it arrives. The company controls part of the roadmap; the market, infrastructure and institutions control another part.

Evidence must therefore be read in relation to the claim it is being asked to support. A useful hierarchy has four levels. Binding evidence — a signed agreement with the party controlling the resource, a final regulatory approval or another enforceable right — can justify a high degree of confidence. Project-specific evidence — a place in a queue, a completed study or a defined scope of required upgrades — establishes that a project has reached an earlier, but still conditional, state. Operational evidence — equipment installed and tested, permits in hand or fuel contracted — demonstrates that the asset can work, but not necessarily that it is entitled to operate at the level or on the timeline assumed. Directional evidence — forecasts, auction results or proposed rules — indicates where the market is heading rather than establishing a right the company actually holds. The analytical error is almost always the same: an investment deck uses level-four evidence to make a level-one claim. In a crowded market, that distinction becomes even more important because the queue, the rule and the relevant standard may all move while the company is executing against them.

One question separates the two faster than any checklist. For this exact site or unit, show me the binding document proving you can operate at the promised level on the promised date, and identify every condition under which that right can be reduced or suspended.

From a diligence snapshot to a living conviction

Traditional diligence is organised around a transaction and therefore tends to produce a snapshot. Deep tech investment unfolds over years, during which the environment surrounding the roadmap can change repeatedly. The more useful model is a living investment conviction: a small number of valuation critical claims are identified at entry, the evidence behind them is made explicit, and the external conditions on which they depend are followed over time. When a regulator introduces an interim category, when a waiting list becomes longer than the company’s build cycle, when established buyers contract years ahead or when a new reporting obligation begins to expose data that did not previously exist, the investor has an early indication that a milestone may no longer mean what it meant at underwriting.

This is where AI-assisted analysis can become genuinely useful. Its value is not in producing another generic diligence report or substituting automated confidence for investment judgement. It lies in continuously reading a fragmented field of regulatory filings, technical notices, company disclosures and market signals, then reconnecting those changes to the few assumptions carrying the roadmap. Used well, such a system can show that a claim has strengthened, that an external dependency has become more fragile, or that the market has begun to redefine the category in which the asset was originally assessed. The objective is not permanent doubt. It is a more durable form of conviction, one that remains connected to the physical, institutional and commercial reality through which deep technologies must eventually scale.

—

About the contributor

Eden Djanashvili runs DeepRadar, an independent practice that checks whether a company’s promises to investors still hold. She reads the contracts and filings behind those promises. When one stops holding, she tells investors before it’s too late to reprice, renegotiate or walk away. She is a jury expert for the European Innovation Council and lectures at ESADE and UPC. console.deepradar.tech

I spent three years commercialising deep-tech inventions and intellectual property in Singapore, followed by two years building a women’s health and medtech venture in the UAE. The technology stacks, founding teams and sectors were different, but the structural contrast between the two ecosystems revealed something consequential that many investment theses neglect: the infrastructure that makes a technology investable is itself a product of geography, and it is distributed profoundly unevenly.

What Singapore taught me: infrastructure precedes investability

By Southeast Asian standards, Singapore is a striking anomaly. It possesses advanced institutional infrastructure that took decades to build and that some of its neighbours have yet to develop. A*STAR provides research funding that can move at a pace approaching commercial relevance. The structures governing spin-offs from the National University of Singapore (NUS) are sufficiently developed to be intelligible to institutional investors. The government is willing to take early, non-dilutive positions in technology companies, helping to mitigate the risk of failure during the pre-revenue phase. English is the operating language of commerce and regulation, supported by a strong and reliable legal system and a stable currency.

Some might attribute these advantages to Singapore’s status as a small, wealthy city-state, but they did not arise automatically. They are the result of deliberate policy choices, decades of calibration and sustained institutional investment. When I founded ScaleHub to commercialise deep-tech IP in Singapore, I was operating within, and benefiting from, that accumulated infrastructure. The research partnerships were complex but governed by frameworks that all parties understood. Grants were not “easy money”; the application processes were demanding, but they were legible. The investors we approached understood and recognised the institutional signals—the NUS provenance and A*STAR endorsement—and adjusted their risk assessments accordingly.

The rest of Southeast Asia presents a different proposition. Countries such as the Philippines, Malaysia, Indonesia and Vietnam each possess significant scientific talent and genuine technological ambition, but often lack some of the connective tissue that converts that ambition into investable ventures. The business environment is generally less frictionless than Singapore’s. Regulatory pathways can be opaque or inconsistent, while government grant programmes, where they exist, frequently operate on timescales that bear little relationship to the capital requirements of an early-stage company. University technology-transfer offices vary enormously in sophistication. The legal infrastructure for equity investment is present, but it is not always calibrated to the specific requirements of deep-tech investing.

The result is that technology emerging from science of comparable quality can face a materially different commercialisation pathway depending on whether it originates in Singapore, Kuala Lumpur, Manila or Jakarta. This is not a judgment on the quality of those ecosystems, but an observation about the current distribution of infrastructure and its implications for investors who prefer to treat geography as a secondary consideration.

What the UAE taught me: ecosystems can be built in real time

The United Arab Emirates presents a different but equally interesting challenge. Unlike Southeast Asia’s relatively mature but uneven landscape, the UAE is constructing an ecosystem at speed, supported by institutional will, sovereign capital and a clarity of strategic intent with few parallels elsewhere.

The result is an unusual commercialisation environment for a deep-tech founder. Infrastructure that took Singapore decades to assemble is being designed and activated in the UAE within a much shorter period. The University of Sharjah and Khalifa University have both been encouraged to demonstrate impact beyond academic publication and are actively seeking commercialisation partnerships in pursuit of that objective. Government-supported institutions such as the Sharjah Research Technology and Innovation Park possess fabrication and prototyping capabilities that would otherwise be difficult or prohibitively expensive for an early-stage company to access, and they are making those capabilities available to qualified startups as part of their wider institutional mandate.

Yuthera, a medtech startup I recently founded, is currently benefiting from partnerships with both universities. When I drove to them unannounced and proposed research partnerships for Yuthera, I was not making as extraordinary a request as it might have appeared. I had arrived at a moment when those institutions were actively looking for founders and companies with whom they could work.

In the UAE, the challenge differs from Singapore’s. The issue is not an absence of infrastructure, but its relative novelty. The strategic intent behind medical-device regulation is clear, but parts of the framework are still being operationalised in practice. The country has sophisticated healthcare systems, yet the pathway from regulatory approval to clinical adoption still requires founders to engage with stakeholders and build relationships through processes that are not fully institutionalised. The investor community is growing, but deep-tech fluency remains concentrated in particular pockets rather than evenly distributed across the ecosystem.

The investment implication: geography belongs in the risk model

For investors, the structural contrast between these geographies leads to a single uncomfortable conclusion: the capital requirement of a given technology cannot be treated as fixed. It is partly a function of how much infrastructure the company must provide for itself because the surrounding ecosystem does not already supply it.

A deep tech company operating in Singapore can often rely on government programmes to absorb some pre-commercial risk, universities to provide research infrastructure and a functional grant system to extend runway and moderate the burn rate. It can also expect institutional investors to recognise the signals those mechanisms generate. The same may not be true for an equivalent company in Manila, Bangkok or Lagos. There, the company may need to replicate much of that infrastructure internally or proceed without it, carrying a burden that is not merely financial but also temporal, organisational and psychological.

Investment theses that treat geography as a secondary variable, evaluating technology as though mature infrastructure were constant across ecosystems, systematically misprice the commercialisation risk of frontier companies operating in infrastructure-sparse environments. This represents both an inefficiency and an opportunity, but only for investors who understand the actual cost of the infrastructure gap and are prepared to provide more than capital to close it.

The connective tissue that we, at Entropia Capital, often describe as unevenly distributed is not simply a market condition to be acknowledged and accepted. Producing that connective tissue is part of the work investors must undertake if they intend to capture value from frontier technologies across Southeast Asia, the Middle East and Africa. The investors best positioned to do so will not be those who have merely learned to tolerate unevenness, but those capable of building the missing infrastructure themselves.

We encounter this question both as investors and as board members. Frontier tech companies are unusually dependent on the quality of their decisions because their principal risks cannot be separated neatly. A technical decision may alter the regulatory pathway; a commercial commitment may impose a new manufacturing architecture; a financing decision may determine which scientific milestones can be reached before the next round. The people closest to each problem rarely possess authority over all of it.

In the early years, the founder holds these elements together. They carry the history of the technology, the reasons behind previous compromises and the relationships on which the company depends. This concentration of knowledge is initially a competitive advantage. If the company grows successfully, however, it can become one of its most important constraints.

Photo by Benjamin Child on Unsplash

The founder becomes the company’s most expensive queue

A founder’s early centrality is usually rational. They possess more context than anyone else, recognise hidden dependencies and can resolve disagreements without convening a constitutional assembly. In technical companies, they may also be among the few people capable of evaluating both the science and the commercial compromise under consideration. The organisation learns to route difficult decisions towards them because doing so produces better answers.

It continues after the volume of decisions has exceeded any individual’s capacity. The founder then becomes a queue: highly intelligent, strategically important and increasingly responsible for the latency of the entire system. Around a board table, this often appears as a discussion about workload. The proposed remedies include better meeting discipline, an executive assistant or the delegation of routine tasks. These measures can provide relief, but they do not address why decisions continue travelling upwards.

Usually, authority has been delegated without being specified. Managers are told to take ownership but remain uncertain which trade-offs they are permitted to make, so they seek approval. The founder, frustrated by the absence of initiative, intervenes and often improves the immediate decision. The organisation learns that escalation remains the safest procedure. A few repetitions are enough to establish the culture. The founder must eventually move from answering the greatest number of important questions to designing how important questions are answered. We have found this to be one of the hardest transitions in company building: it sounds like a promotion, but often feels like a loss of control.

Decision-making needs an architecture

Pushing decisions downwards does not mean treating every decision alike. Choices that alter the technical architecture, capital structure, regulatory exposure or survival of the company deserve senior attention. Reversible operating decisions generally need speed. A scaling organisation should make the distinction explicit: where does authority sit, when is consultation required, and what conditions trigger escalation? Without this architecture, “empowerment” remains an attractive word attached to a system of informal centralisation.

This is especially important in frontier tech companies, where decisions regularly cross functional boundaries. Engineering may understand what can be built, regulatory may understand what can be approved, commercial may understand what customers will buy, and finance may understand what the company can afford. None possesses the complete answer. The objective is not to remove the founder from consequential decisions but to ensure that the organisation can integrate these perspectives without asking the founder to arbitrate every disagreement.

A board can help by examining the flow of decisions rather than only their outcomes. Which decisions repeatedly arrive late? Which ones return to the founder after apparently being delegated? Where do teams wait because several people can object but nobody can decide? These questions reveal more about operating capacity than another discussion of organisational charts. Authority does not exist because a box appears beneath someone’s name. It exists when the organisation knows which decisions that person can make and expects those decisions to stand.

Reality deteriorates as it travels upwards

At twenty people, a founder can observe much of the company directly. At two hundred, they observe representations of it. Investors and boards are one step further removed. Information arrives through dashboards, management meetings, board materials and conversations that have already passed through several layers of interpretation. No deliberate deception is required. Each layer removes some uncertainty, sharpens the explanation and makes the situation slightly more suitable for presentation.

Good news travels rapidly because it is easy to deliver. Bad news waits for confirmation. A technical delay remains manageable until the recovery plan fails; a customer concern remains anecdotal until procurement stops responding; a hiring problem remains temporary until the preferred candidate joins a competitor. Hope, indispensable to entrepreneurship, begins performing tasks for which evidence would be more suitable. This matters particularly in frontier technology, where technical uncertainty cannot be managed through optimism. Biology remains indifferent to investor updates. Manufacturing yield does not improve because the quarterly narrative requires it. Regulators have yet to adopt the venture industry’s preferred relationship with deadlines.

As board members, we contribute to this filtration through our own reactions. If the person raising a problem is treated as its source, the organisation learns quickly. Future problems will arrive with more context, better formatting and less time remaining to solve them. We therefore try to distinguish bad outcomes from bad operating behaviour. Bad news raised early, accompanied by serious analysis and a clear owner, indicates that the information system works. Bad news concealed, minimised or repeatedly presented without ownership indicates that it does not. A board should react very differently to the two. Reality eventually enters the room; the only variable is how much time and capital have been spent before it does.

A dashboard is not an operating system

Scaling companies tend to acquire metrics at roughly the same time they acquire managers. Each manager needs information, each function needs objectives, and the board would reasonably prefer not to govern by anecdote. The result can be a large quantity of increasingly precise data describing an organisation that remains difficult to control. We do not need every board pack to contain more information; we need it to make the essential information harder to avoid.

Are the few trajectory-defining outcomes moving? What currently prevents them from moving? Who owns the constraint? Which decision is required, and from whom? Useful visibility does not attempt to create a perfect representation of the company. It shortens the distance between reality and action. A technically comprehensive dashboard can conceal as much as it reveals if the central risk appears on page forty-seven, surrounded by measures that are improving.

The same standard applies to meetings. A recurring meeting should make decisions, allocate resources, remove constraints or review commitments. The discipline is almost embarrassingly simple: meetings end with decisions, owners and dates, and the next meeting begins with the commitments made previously. Actions should not disappear into minutes that demonstrate excellent administrative hygiene but no institutional memory. Accountability is created less by intensity than by recurrence. What matters is not the dramatic intervention after a missed milestone, but the expectation that every commitment will return to the room.

What we believe an investor can usefully do

The venture industry has stretched the expression “hands-on” until it encompasses almost every form of investor behaviour short of remaining entirely absent. At Entropia Capital, operational involvement has a more specific meaning. Frontier tech companies’ principal constraints often span science, governance, regulation, industrial execution, capital formation and geography. Solving one component in isolation may not alter the company’s trajectory.

A European technology may require Asian manufacturing capability, Gulf infrastructure capital or North American commercial access. A technically successful company may remain unfinanceable because its governance, leadership structure or regulatory sequencing is unintelligible to the investors required for the next stage. An apparently substantial commercial opportunity may be strategically wrong because the company cannot deliver it without fragmenting its core platform. Our role is not to occupy the founder’s chair from a safer distance, nor to provide a collection of introductions and describe the resulting email traffic as value creation. We try to identify the transition that matters, assemble the relevant resources around it and remain involved until the company has materially changed state.

Sometimes this means assisting with senior recruitment or governance. Sometimes it means restructuring the financing logic around the real technical milestones. Sometimes it means connecting European science with Asian industrial capability, Gulf capital or American commercial access. And sometimes it means arguing against an attractive expansion because the underlying system cannot yet absorb it. The work varies because the constraints vary, but the principle does not: our involvement should make the company more capable and, eventually, less dependent on us. This is what we mean when we describe ourselves as blue-collar investors and operators. The phrase implies work, proximity to the constraint and the mildly inconvenient possibility of being judged by a visible result.

Building an institution

A company becomes an institution when the quality of its decisions no longer depends on one person being present in every consequential conversation, and when reality can travel upwards without being made comfortable along the way. The founder remains essential, but their leverage increasingly comes from designing the system rather than operating every part of it.

The board has a role in that transition. It should help make choices clearer, bad news safer to disclose and commitments harder to forget. It should bring perspective without creating another layer of management, and support the founder without preserving a form of dependence that the company has already outgrown.

At Entropia Capital, we believe the same test applies to the investor. Our work should leave behind better decisions, clearer ownership and greater operating capability, not simply a longer list of meetings in which we participated. Capital can finance the organisation. Building the institution requires something more demanding: an honest view of reality, a deliberate architecture for decisions and the discipline to make both survive beyond the people who created them.

This should be the point at which the company accelerates. Sometimes it is. Yet, sitting on boards and working alongside founders, we have repeatedly seen a more complicated reality: capital accelerates not only a company’s strengths, but every unresolved ambiguity inside it.

Unclear priorities acquire teams, weak decisions acquire budgets and temporary workarounds become departments. A founder who was already involved in too many questions becomes involved in a greater number of more consequential ones. The organisation becomes larger without necessarily becoming more capable.

The VC industry prefers a pleasantly linear account of company building: a startup creates something valuable, demonstrates that a market exists, raises a substantial round and uses the proceeds to scale. Headcount increases, commercial activity expands, and a photograph is taken in which the founders appear tired but optimistic beside a conspicuously large number. The future, having now been financed, is expected to arrive more quickly. Our experience is that the passage from startup to institution is considerably less linear. This is not an argument against growth, hiring or large financing rounds. Frontier tech companies cannot industrialise through restraint alone. Labs, certification programmes, manufacturing systems and global commercial teams have the inconvenient habit of requiring actual money. It is an argument against confusing the accumulation of resources with the construction of an institution.

Capital changes the internal economy

Scarcity performs several management functions rather efficiently. When a company has four to six months of cash, a small engineering team and a product that does not yet work reliably, prioritisation tends to be direct. People may disagree about the solution, but the problem itself is difficult to avoid. The constraint is visible, the responsible people are known and decisions are made with an intimacy that no management framework has yet improved upon.

A large financing round changes this economy. Several previously impossible projects become possible at once. New geographies appear accessible, long-deferred hires can be made, customers request adaptations and industrial partners propose joint initiatives. The board, having approved an ambitious plan, understandably expects to see ambition occur. In board discussions, we often encounter projects that are perfectly defensible when considered individually. The new geography is attractive. The second application has a large addressable market. The strategic partnership brings a respected name. The senior hire appears capable of building the organisation required for the next stage. The difficulty emerges when these decisions are considered together: what looks like a series of rational initiatives can amount to an incoherent company.

Strategy is no longer enforced by the impossibility of doing everything. It must now be imposed deliberately by management, at precisely the moment when saying no has become politically and psychologically more difficult. An opportunity accompanied by a famous customer logo rarely presents itself as a distraction. Each arrives with a credible argument, an internal sponsor and a spreadsheet showing considerable upside. The spreadsheet may even be correct in the narrow sense. What it omits is the organisational cost of pursuing several correct things simultaneously. In deep tech, that cost becomes physical: a new application may require different engineering specifications, suppliers, qualification processes, regulatory pathways and commercial expertise. The company has not added a feature; it has begun constructing another company without completing the first one.

Strategy begins with subtraction

A priority is something that takes precedence over something else. Corporate language has been remarkably successful in separating the word from this meaning. We regularly see scale-ups with six or eight strategic priorities, several annual objectives beneath each one and functional roadmaps bearing only a diplomatic relationship to either. Everyone is busy and progress is extensively documented, yet the central constraint survives from one quarter to the next.

When we work with a board or management team, we try to identify the small number of outcomes that could materially alter the company’s trajectory. In frontier technology, these are rarely generic ambitions such as “accelerating commercial growth” or “building operational excellence.” They are concrete changes in the company’s condition: moving a process from laboratory reproducibility to a qualified production line; obtaining the certification without which customers cannot deploy the product; proving that unit economics remain credible outside a subsidised pilot; or securing the reference customer that changes how the market prices technical risk.

The distinction between activity and outcome sounds elementary because it is. It is also violated with impressive consistency. Hiring a commercial team is an activity; establishing repeatable sales in a defined segment is an outcome. Building a manufacturing facility is an activity; producing at the required yield and cost is an outcome. Signing a memorandum with a large corporation is an activity—and occasionally a form of corporate theatre. Converting it into a paid deployment capable of surviving procurement is an outcome. Once the few outcomes that matter are clear, management can place everything else below the line. “Not now” remains one of the most useful strategic positions available to a company, even if it lacks the emotional satisfaction of a launch announcement.

Outcomes need owners

As companies grow, ownership tends to migrate from people to nouns. Engineering owns the product, operations owns manufacturing, business development owns the partnership and the leadership team owns the strategy. A steering committee may then be introduced to own whatever remains insufficiently owned by the first four. Functional responsibility is necessary, but important company outcomes are rarely functional. An industrial ramp may depend simultaneously on design engineering, procurement, quality assurance, recruitment, customer acceptance and financing. Each function can perform its assigned work competently while the milestone continues to slip.

We have seen this pattern in many board meetings. The presentation is detailed, every delay has an explanation and no individual statement is necessarily incorrect. Engineering explains that procurement arrived late; procurement explains that specifications changed; commercial explains that the customer introduced a new requirement. Everyone is factually right, and the company still misses the milestone. The problem is not necessarily insufficient collaboration. It is the absence of a person whose responsibility survives the boundaries between functions.

Every trajectory-defining outcome should have one identifiable owner. That person need not have hierarchical authority over everyone involved, but must have the decision rights, information and institutional backing required to resolve trade-offs. Above all, the organisation must understand that this person is accountable for the result, not merely for producing updates about it. Single ownership is not a moral judgment; it is an information architecture that obliges one person to maintain a complete view of the problem even when its components sit elsewhere.

Headcount is an expensive form of optimism

When a team falls behind, the need for additional people can appear self-evident: there is too much work and too little capacity. Hiring seems less like a strategic choice than the recognition of arithmetic. Sometimes the arithmetic is real. A new production line requires operators, regulatory submissions require specialised expertise, and a company entering a market cannot indefinitely substitute founder travel for local commercial capability.

In our experience, however, headcount is also the most socially acceptable explanation for organisational underperformance. It places the problem in the future, i.e, once the right people arrive, rather than in the company’s present design. Adding people to unclear ownership does not create clarity; it creates more interfaces through which clarity must travel. Adding managers to a slow decision system can make decisions slower with greater professionalism. Adding a programme office to ten competing priorities may improve the reporting of their competition.

When a hiring plan reaches the board, we want to understand more than how busy the team has become. What constraint will this person remove? Why can it not be removed through a decision, a change in sequence, a narrower objective or the elimination of unnecessary work? What measurable difference should exist once the person is in place? This is not an argument for permanent understaffing, another fashionable way of exhausting competent people. It is an argument for treating headcount as the consequence of an operating design rather than as the design itself.

Scaling is not enlargement

A company has not scaled merely because it employs more people, occupies a larger facility or has raised a later letter of the alphabet. It has scaled when it can produce important outcomes repeatedly without requiring a proportionate increase in founder attention, organisational friction or capital consumption, i.e., when its capabilities expand faster than the complexity created by that expansion.

This is a high standard, particularly for frontier companies. Their products must operate in the physical world, where deployment encounters supply chains, industrial standards, clinical evidence, infrastructure and customers generally less impressed by the pitch deck than the pitch deck anticipated. That is precisely why the organisational question matters. Scientific advantage creates an opportunity and capital creates capacity, but neither automatically creates an institution. That requires choosing fewer outcomes, assigning genuine ownership, hiring against constraints and building a company capable of making decisions without routing every difficult question through the founder.

Capital gives a business more available paths. The responsibility of leadership is not to travel several of them enthusiastically, but to determine which one leads to an enduring company and close the others for long enough to matter.

One of the most useful things we can do at Entropia is therefore also one of the least glamorous: help a company subtract. The financing round may accelerate the journey. It does not choose the destination or build the vehicle.

He has simply run out of road.

For several months, we kept returning to the same problem. His performance remained strong, perhaps stronger than before, but the internal logic that had once sustained it was disappearing. The company still mattered to him. His role no longer did. Then the other day he suggested what appeared to be a practical solution. “Perhaps I should ask for a raise. It has been a while since I progressed. Maybe that is what is missing.” I asked whether he genuinely believed another 15% would make him want the job again. It was not a particularly sophisticated intervention. It worked because he already knew that it would not.

The question was never whether he deserved more money. He probably did. The question was whether money and motivation had begun to stand in for one another because neither he nor the company had found a better language for what was happening. What he wanted was not primarily an improved version of the same role. He wanted another cycle: a different order of problem, a new source of uncertainty, perhaps eventually a company of his own. The salary request was merely the first negotiable expression of something that was not yet negotiable.

Once he saw this in himself, he began to recognise it in his team. Several of his commercial leaders had recently asked for higher compensation while appearing less engaged than ever. He had assumed that they wanted more money. Some certainly did. Others were asking whether the company still contained a future in which they could recognise themselves. This is a distinction that early-stage companies tend to discover too late.

The ambiguity of compensation

The claim that money does not motivate people is plainly untenable. Compensation affects who joins, who remains and what sacrifices a company can reasonably ask someone to make. It determines material security and opportunity cost. It also conveys information. A salary is not only purchasing power; it is an institutional judgment about value.

Underpayment therefore creates more than dissatisfaction. It introduces a fracture between the company’s account of someone’s importance and the evidence contained in their payslip. Mission, equity and proximity to charismatic founders do not erase that fracture. In some cases, they make it worse by turning commitment into a resource the company quietly consumes.

Yet the inverse proposition, i.e, that higher compensation can restore durable motivation, is equally weak. A few years back, a large meta-analysis found only a modest relationship between pay level and overall job satisfaction. The correlation was approximately 0.15, rising to around 0.23 when the question was narrowed to satisfaction with pay itself. More money can make an employee happier with their compensation without materially changing their relationship with the work. More recently, Gallup’s retention data reveal a similar ambiguity. Pay and benefits were the largest single reason American employees gave for leaving in 2024, but accounted for only 16% of departures. Engagement and culture represented 37%; wellbeing and work-life balance, another 31%. None of this makes compensation secondary. It makes compensation multidimensional.

When someone asks for a raise, they may be pricing their labour. They may also be pricing an accumulated injustice, a loss of status, a repetitive role or the growing probability that they will leave. The same number can contain several different claims.

Organizations generally respond to the number because the number is administratively convenient. It can be benchmarked, approved, deferred or rejected. The underlying question, i.e., whether the employee still has a viable future inside the company, is much harder to place on a compensation committee’s agenda.

The frontier tech talent problem

This ambiguity is particularly consequential for the companies Entropia Capital backs and builds. Frontier tech businesses are unusually dependent on concentrated knowledge. A small number of people often carry a disproportionate share of the company’s technical memory, regulatory understanding, commercial trust and operating judgment. Their value is not fully contained in a job description. It resides in what they have learned over time: why one technical path failed; which data can be trusted; which clinical, scientific or industrial partner can execute; where a regulatory argument remains fragile; which customer objection conceals a real implementation constraint; how the system behaves outside a controlled demonstration.

In conventional labor accounting, an employee leaves and another is recruited. In a deep tech company, the departing employee may take part of the company’s accumulated learning curve with them. This risk is intensified in the small, open economies where Entropia operates. Singapore, Dubai and Luxembourg can attract exceptional international talent, but their specialized labor markets remain relatively thin. Expertise is mobile, replacement cycles are long and senior hires often carry significant geographic and family constraints. A hiring error cannot always be absorbed by a sufficiently large local market.

Retention in these companies is therefore not merely an HR objective. It is a form of continuity engineering. Gallup estimates that replacing an employee in a technical role can cost around 80% of annual salary, while replacing a leader or manager may cost approximately twice their salary. Such estimates are necessarily imprecise, but they still omit the effects that matter most in an early-stage company: delayed milestones, weakened customer relationships, slowed regulatory work and the departure of knowledge that was never documented. For a company with limited runway, losing a critical operator can alter the financing trajectory.

The usual response is to pay more. Sometimes this is correct. It is rarely sufficient as a theory of retention.

When competence becomes a trap

The people most exposed to this problem are often the company’s strongest performers. They join during a phase of genuine uncertainty. The initial role is poorly defined because the company itself is poorly defined. They build the function, recruit the first team, close the first serious customer or convert a scientific proposition into an operational system. The work is exhausting, but it is formative. Every quarter changes the nature of the problem.

Then the company begins to stabilize. What was once invention becomes execution. The employee’s judgment is encoded into routines, meetings and processes. Because they are exceptionally reliable, they are given more responsibility for maintaining what they created. From the company’s perspective, this is evidence of trust. From the employee’s perspective, it can feel like being imprisoned by competence. They are too valuable in the current role to be allowed to abandon it, but no longer sufficiently challenged by the role to inhabit it fully.

Startups often misread this moment. They add a title, a team or a performance bonus. Yet the problem is not necessarily a shortage of volume. It is the disappearance of novelty. More of the same responsibility is not a new cycle. A larger team is not inherently a larger problem. A promotion that leaves decision rights unchanged merely formalizes the existing constraint. Research on self-determination at work is useful here, provided it is not reduced to a motivational slogan. It suggests that sustained motivation depends in part on autonomy, developing competence and meaningful connection to others. Challenge matters when it expands agency and capability. It matters less when it simply increases load. A top performer asking for a raise may therefore be asking whether the company can still produce a version of them that does not yet exist. That is not a question compensation can answer alone.

Recognition is not praise

There is another recurrent source of confusion. Companies often respond to the limits of compensation by invoking recognition, as though a sincere thank-you were the non-monetary substitute for a salary adjustment. Recognition is not a substitute for fair pay. Nor is it synonymous with praise. In organizational terms, recognition means that contribution changes position. It produces greater trust, access, authority, visibility or scope. It alters the relationship between what the employee carries and what the institution allows them to decide.

Gallup and Workhuman followed nearly 3,500 employees from 2022 to 2024 and found that those receiving high-quality recognition were 45% less likely to have changed organizations two years later. Only 22% of employees believed they received the right amount of recognition. The finding should not encourage startups to industrialize gratitude. Recognition systems become hollow when they increase the frequency of appreciation while leaving the distribution of power untouched. A person who repeatedly rescues important projects but remains excluded from strategic decisions is not suffering from insufficient praise. They are receiving contradictory info: the company depends on their judgment operationally but does not recognize it institutionally. A bonus may soften that contradiction. It does not resolve it.

The company they joined no longer exists

There is a further difficulty that appears as startups mature. Employees can remain deeply loyal to the company they joined while becoming alienated from the company that now exists. This is not always a story of cultural decline. Transformation is part of company building. A research project must become a product; a product must develop commercial discipline; an informal team must acquire governance; technical possibility must submit to procurement, regulation and economics. The early employees who made one stage possible are not automatically the people who will find the next stage meaningful. Some flourish as the organization scales. Others discover that their commitment was attached to a particular kind of company: smaller, more technical, more direct, less procedural or more willing to operate without certainty. At this point, compensation acquires another function. It becomes the price of estrangement. The employee is no longer being paid only for the work. They are being paid to remain inside an organization whose direction, tempo or politics no longer fit them. A sufficiently large increase may delay the departure. It may even be economically rational for both parties. But the company should be precise about what it has purchased. It has bought time, not necessarily renewed commitment.

Counteroffers often fail because they answer the question the company can afford rather than the question the employee is asking. “If we matched the offer, what would still be wrong on Monday morning?” is generally more revealing than “What would it take to keep you?”

The two conversations

When a critical employee asks for higher compensation, the company owes them two separate conversations.

The first concerns fairness. Is the employee paid appropriately relative to the external market, their internal peers and the responsibility they actually carry? Has the role expanded without a corresponding change in compensation? Is the equity meaningful, or is it being presented as valuable while its probability, dilution and liquidity remain obscure? This conversation should be conducted without appealing to culture or purpose. If the person is underpaid, the company should say so and, where possible, correct it.

The second conversation concerns trajectory. Does the employee still want the future the company can realistically offer? Are they learning? Do they possess authority commensurate with their responsibility? Do they identify with what the company is becoming? Is there another cycle available inside the organization, or would creating one amount to inventing a position that neither the company nor the employee truly needs? These conversations should remain separate because each can be used to avoid the other.

Managers sometimes psychologize a legitimate compensation request: perhaps the employee needs more recognition, more purpose or a development plan. Employees, in turn, sometimes financialize a deeper departure: perhaps another 15% will make a completed chapter feel unfinished again. The purpose is not to discover whether money or motivation is the “real” issue. Both can be real. The purpose is to prevent one from concealing the other.

The problem is that most companies begin this inquiry only after the salary request has been formalized. By then, the employee has assembled market data, spoken with peers or taken calls from recruiters. A diffuse dissatisfaction has become a position to defend. The better conversation takes place earlier. Not “Are you happy?” (a question too broad to produce useful information) but something closer to:

What would need to be true a year from now for you to believe that staying was the right decision?

The value of the question lies in its time horizon. It forces the company and the employee to consider not only the quality of the present arrangement, but whether that arrangement contains a credible future.

Retention is not permanence

Venture-backed companies tend to treat retention as an unconditional good. In reality, permanence is neither possible nor desirable. The objective is not to keep every strong employee indefinitely. It is to preserve alignment between the company’s next constraint and the employee’s next ambition.

Sometimes that alignment can be constructed. A new market, product, technical problem or level of authority can create a genuine second cycle. At other times, the company has no honest way to offer what the employee needs. Manufacturing a role to prevent a departure can be more expensive than the departure itself. It creates ambiguous authority, slows succession and places a disengaged senior person between the company and someone for whom the next stage would be formative.

In such cases, helping an exceptional employee leave well may be a stronger act of management than retaining them through a counteroffer. This is particularly difficult in frontier technology, where talent scarcity encourages companies to hold on to people long after the underlying exchange has weakened. But physical presence is not the same as commitment, and continuity cannot be secured indefinitely through compensation. A raise can correct an imbalance. It can acknowledge an expanded role, repair a legitimate grievance or share more fairly in the value someone helped create. These are sufficient reasons to grant one.

What it cannot do, by itself, is restore a future.

Before deciding what someone should be paid to stay, a company should understand what, precisely, they are being asked to stay for.

This distinction matters because the venture industry spends an extraordinary amount of time debating the price of startups and remarkably little time debating the price of venture capital itself. We argue over whether a company is worth $XX million or $XX+5 million while accepting, with near-liturgical consistency, that the vehicle holding it should conform to some variation of 2/20. The underlying assets may range from a preclinical therapeutics platform in Paris to a robotics company in San Francisco or a crypto business in Dubai. Their capital requirements, timelines and risk structures bear little resemblance to one another. Yet the wrapper remains suspiciously uniform.

At Entropia Capital, we decided to question the wrapper a long time ago, not because the traditional fund structure is inherently defective, but because it was designed for a different investment logic from ours. We invest in frontier companies whose central difficulty is rarely technological risk alone. Their problems sit at the intersection of science, governance, regulation, infrastructure, capital formation and international execution. They are systems businesses, and systems businesses are poorly served by capital that behaves as though the relevant work ends when the wire transfer is completed.

Our response is built around three principles: no recurring management fees, substantial operational involvement, and the active creation of liquidity opportunities earlier in the company’s development. The precise economics can vary with the mandate. The principles do not.

The wrapper eventually shapes its contents

Management fees perform a legitimate function. They finance diligence, reporting, compliance and the institutional machinery required to manage other people’s capital. Serious investing cannot be sustained indefinitely on optimism, deferred compensation and airport coffee.

Yet fee structures also produce organizational gravity. A larger fund supports a larger team; the larger team requires more assets; more assets require larger investments; and the strategy gradually adapts to the institution created to execute it. Fundraising success becomes economically tangible today, while investment success remains provisional, sometimes for a decade. This tension is particularly acute for smaller funds. On a $25 million vehicle, a conventional 2% annual fee represents $500K before legal costs, administration, travel and the expense of maintaining a genuinely international operation. It is enough to reduce the capital available for deployment, but rarely enough to support the institutional platform described in the presentation. Small managers can therefore find themselves in an awkward middle ground: too expensive to function as pure investment partnerships, too small to reproduce the infrastructure of established firms, and under pressure to raise a successor vehicle before the first has returned meaningful capital.

We chose a different constraint. Entropia does not rely on recurring management fees; the investment capital remains as productive as possible, while the operating platform must earn its own living. We don’t see that as a discount, more as a discipline.

An operating company with investment capital inside it

Without recurring fees, operational capability cannot remain a promise subsidized by LP commitments. It must produce revenue, develop an asset or demonstrably improve an investment. Entropiatherefore operates through four connected activities: investment, venture building, venture architecture and ecosystem architecture.

The investment activity provides exposure to asymmetric outcomes in deep tech and frontier industries. Venture building creates companies where fragmented markets do not reliably generate investable ventures on their own. Venture architecture addresses governance, international structuring, strategic finance, recruitment, regulatory positioning and commercial development. Ecosystem architecture uses executive education to build knowledge, relationships and entrepreneurial density around the markets in which we operate.

These are sometimes called “platform services” in venture capital, which can mean anything from a genuine operating capability to a well-designed page on a website. In our case, they must support themselves economically while reinforcing the investment activity. This is less a fund surrounded by services than an operating company with investment capital embedded inside it. The architecture is deliberately uncomfortable. Comfort, in asset management, is not necessarily an investor benefit.

Deep tech operates on several clocks

Deep tech is commonly described as requiring patient capital. This is correct, but incomplete. Scientific and industrial development can be slow. Regulatory approvals take time; manufacturing systems do not scale at software speed merely because their founders have adopted software vocabulary. Biology, in particular, remains wonderfully indifferent to quarterly reporting. Yet a long company-building cycle does not mean every investor must own every share until the final exit.

A frontier company passes through several states of legibility. At inception, it may combine exceptional science with an incomplete team, uncertain regulatory sequencing and a corporate structure assembled by people who have not yet encountered an institutional investor. Several years later, the technology may still be pre-scale, but the company can possess protected intellectual property, credible governance, early customer validation, a coherent financing architecture and leadership capable of absorbing substantially more capital.

The technology remains young. The investment risk has nevertheless changed. We describe this transition as becoming venture-proven before becoming fully market-proven. It is the point at which a technically compelling but institutionally difficult company becomes intelligible to larger venture funds or strategic investors. Our work is concentrated around that transition. We help strengthen leadership, clarify positioning, structure governance, navigate regulatory pathways and connect companies to the geographies most relevant to their development. Depending on the business, European science may need to be combined with Asian manufacturing, Gulf infrastructure capital or American commercialization.

These interventions are not peripheral to deep-tech investing. They determine whether the technology receives sufficient capital and time to mature. They can also create an earlier liquidity boundary.

Ownership duration is not technology duration

Traditional venture often collapses technological duration and ownership duration into a single assumption: invest early, follow the company, and wait for an acquisition or IPO. We prefer to separate the two. Once a company has crossed an important threshold of institutional credibility, some early ownership can potentially be transferred to an investor better equipped for the next stage. This may occur through a strategic transaction, a financing round with a secondary component, a later-stage fund or a partial sale to a corporate partner.

From portfolio value to returned capital

The venture industry speaks reverently about power laws and rather awkwardly about distributions. Paper value is treated as performance with a time delay. Sometimes the delay lasts longer than an undergraduate education, a doctorate and the first several years of the resulting academic career. Yet DPI remains a refreshingly physical metric: capital was invested, and capital came back. IRR adds the equally impolite question of when.

Fragmentation as an investable inefficiency

The model is inseparable from where we invest. Technological capability has globalized faster than venture infrastructure. Exceptional researchers and founders now emerge across Singapore, Paris, Munich, Dubai, Bangalore, Seoul and San Francisco. What remains unevenly distributed is the connective tissue: institutional trust, specialized capital, regulatory fluency, commercial access and cross-border execution. Many frontier companies are overlooked not because their technology is weak, but because they are difficult to parse.

A different specification for the venture product

Rharbaoui argues that a fund is ultimately an expression of its people, strategy, access and accumulated performance. We agree. Venture remains intensely intuitu personae: LPs underwrite individuals exercising judgment in conditions where the data are incomplete and the future stubbornly refuses to resemble a spreadsheet. But if the manager is the product, the economic structure is its specification.

Our specification is relatively simple. Preserve the productivity of investment capital. Sustain the operating platform through activities that create independently recognizable value. Tie the economics primarily to outcomes. Treat liquidity as part of portfolio construction rather than an administrative event at the end of it. The details need not be identical in every situation. A passive minority investment, an intensive company-building mandate and a cross-border restructuring are not the same product; pretending otherwise would replace thoughtful alignment with contractual symmetry.

The model should consequently be judged against three questions:

Does the absence of recurring fees result in more capital reaching companies? Does operational involvement create measurable changes in their quality and investability? Does the portfolio return capital earlier and more consistently than a comparable early-stage strategy?

If the answers are no, the wrapper has failed, whatever its intellectual elegance.

We wanted Entropia Capital to share more of the entrepreneurial condition: to earn revenue before comfort, remain close to execution, create assets rather than accumulate overhead and depend disproportionately on the value we help produce. This is why we describe ourselves as blue-collar investors and operators. The phrase is intentionally inelegant. It implies showing up, performing work that is visible in the result, and accepting that capital alone does not confer usefulness.

The future of venture capital will not be defined by a universal replacement for 2/20. Large institutional funds, specialist partnerships, evergreen vehicles, venture studios and operator-investor platforms will coexist because different risks require different wrappers. But wrappers should express strategies rather than conceal their contradictions.

Ours expresses a particular thesis: investment capital should remain productive; operational capability should justify itself; economics should reflect actual contribution; deep-tech ownership should be managed across distinct risk stages; and geographic fragmentation can be converted from a founder’s burden into an investor’s edge.

A venture fund is indeed a financial product. Unlike an option, however, it cannot be priced from volatility alone. Its value depends on whether the people inside the wrapper can change the trajectory of the underlying assets, and whether, at some point, they remember to return the money.

In the U.S., this tolerance for inefficiency is not cultural indulgence but structural consequence. A single vintage of venture capital can produce hundreds of venture-scale companies, a deep secondary market for talent, and repeated exit cycles within a decade. Failed founders are recycled into new ventures, employees redeploy laterally, and capital is continuously repriced through public markets and late-stage secondaries. Losses dissipate because the system is thick.

Small ecosystems as institutional environments

Small (eco)systems do not enjoy this luxury. The places where Entropia Capital works in Singapore, Dubai, and Luxembourg are not reduced versions of larger ecosystems; they are structurally different environments where venture capital operates less as a market activity than as institutional infrastructure. Talent is imported rather than endogenous, capital is mobile rather than anchored, and exits are predominantly external.

In such systems, venture capital is not merely a participant in innovation, but instead a coordinating mechanism whose failures compound rather than dissipate. The last decade, marked by abundant liquidity and the global diffusion of the “founder-friendly” service model, revealed a critical mismatch: a model designed for abundance was applied to scarcity-constrained systems, producing fragility instead of resilience.

The later emphasis on Singapore reflects its position on a longer innovation arc: roughly thirty years of deliberate policy construction, compared with a decade of accelerated experimentation in the Middle East and a much earlier, finance-centric development path in Luxembourg.

The rise of the service model

The service model, born in the aftermath of the 2008 financial crisis and institutionalized during the era of zero interest rates, reframed venture capital as a form of companionship rather than selection.

Funds increasingly competed not on capital allocation discipline, but on relational proximity to founders. Firms such as Andreessen Horowitz institutionalised the platform model at a moment when U.S. markets were simultaneously deepening (cloud, mobile, fintech) and accelerating. Their services reduced coordination costs in environments where hiring velocity, regulatory exposure, and media narratives were themselves system-level constraints. In deep, liquid ecosystems, this reorientation translated into structurally weaker capital efficiency, i.e., often yielding mediocre risk-adjusted returns for those mimicking the Tier 1 funds without the firepower, while remaining masked by scale, liquidity, and abundant exit optionality.

In small ecosystems, it tends to produce social capture. When everyone knows everyone, governance quickly becomes reputational rather than functional. Boards turn into social equilibria where confrontation is deferred, underperformance is normalized, and capital continues to flow not because probability warrants it, but because relationships do.

Singapore offers a telling illustration: a highly programmatic innovation state, efficient at grant allocation and infrastructure, but ill-equipped to tolerate prolonged private-sector delusion. Over there, service-oriented venture capital, imported wholesale, softened the very discipline the system required most, i.e, leading to companies that survived on signaling and public support long after private markets would have forced correction. It was not naive imitation; it was a rational response to structural gaps. A young startup ecosystem, limited domestic talent pools, and the absence of serial founders meant that early-stage companies required scaffolding that markets could not yet supply. Platform teams, government-adjacent accelerators, and co-investment schemes functioned as ecosystem primers, accelerating time-to-formation rather than time-to-exit.

It is not uncommon in Singapore to encounter Series A or B companies with modest commercial traction, extended public-sector support, and stable venture backing over longer horizons than those typically observed in large private markets. This pattern is best understood not as underperformance, but as a locally coherent equilibrium. When founders prioritize regional consolidation, institutional partnerships, and measured internationalization, persistence becomes a rational outcome. In such settings, continuity of capital and support functions as a stabilizing mechanism while companies mature within a constrained but highly structured environment.

This behavior reflects the specific architecture of Singapore’s ecosystem. The system is particularly effective at company formation, early validation, and institutional coordination, while the pathways to large, absorptive end-markets remain comparatively selective and resource-intensive. International expansion (toward the U.S., China, or major emerging markets) often requires deliberate, stepwise engagement rather than rapid scaling. As a result, many companies optimize for durability and optionality before pursuing aggressive market transitions. The outcome is not stagnation, but a portfolio of enterprises evolving along longer, more deliberate trajectories.

Within this context, the development of service-oriented venture capital can be seen as an adaptive response. In a frictional and still-maturing ecosystem, bundled support, i.e., spanning talent access, regulatory navigation, signaling, and ecosystem coordination, helps reduce early execution risk and compensates for missing market infrastructure. Such models are particularly effective in the formative stages of an ecosystem, where learning curves are steep and institutional interfaces complex. They contribute to system coherence by aligning founders, investors, and public stakeholders around shared operational norms.

As the ecosystem evolves, however, the opportunity shifts toward refining the capital narrative that accompanies this support. Equity structures, financing expectations, and liquidity pathways increasingly benefit from calibration to the realities of constrained domestic markets and multi-step internationalization. Rather than mirroring the timelines and valuation logic of large markets, later-stage companies in Singapore invite models that recognize extended value formation, earlier partial liquidity, and differentiated risk-return profiles aligned with regional and cross-border growth patterns.

Seen through this lens, Singapore’s venture landscape is not a story of misalignment, but of progression. The next phase lies in articulating financing frameworks that explicitly acknowledge the ecosystem’s strengths, i.e., discipline, coordination, and durability, while accommodating the longer arcs required to bridge toward global scale. In small systems, venture capital matures not by abandoning support, but by pairing it with capital structures and equity narratives designed for continuity, optionality, and long-term institutional resilience.

In Dubai, the dominant risk is not over-programming but over-velocity. Capital arrives quickly, often tied to thematic waves (e.g, crypto, Web3, AI) while governance norms reset with each cycle. Founder reverence, combined with rapid deployment, produces companies optimized for narrative timing rather than operational depth.

Luxembourg illustrates the inverse problem: an ecosystem optimized for capital movement rather than company formation. The jurisdiction excels at structuring—fund vehicles, holding companies, secondary transactions—but lacks the entrepreneurial density to sustain narrative-led venture building. The result is often immaculate architecture surrounding insufficient operational mass.

Capital engineering as institutional design

What emerges from these environments is not a call for harsher venture capital, but for more mechanical one. As such, the concept of capital engineering is not a stylistic correction; it is an institutional necessity in small (eco)systems.

Its first principle is that liquidity cannot remain a distant, binary event. Where founders are often expatriates and senior talent is globally benchmarked, early and partial liquidity is not

indulgence. It is retention infrastructure. Recurring tender offers, long treated as exceptional, function in small ecosystems as behavioral stabilizers. They reduce pathological risk-taking driven by personal financial exposure and introduce regular pricing events that puncture internal valuation myths.

In Singapore, tender mechanisms increasingly determine whether second-generation founders remain in the ecosystem or return capital (and themselves) abroad. In Dubai, they counterbalance volatility by anchoring incentives over time. In Luxembourg, they align naturally with a jurisdiction already optimized for cross-border capital flows and secondary transactions. Liquidity, in this framework, is not the end of the venture journey; it is a tool for governing risk along the way.

The second principle of capital engineering is unbundling. The historical bundling of capital, advice, and execution was a convenience, not an economic truth. In small systems, it becomes a distortion. When advice is paid for with equity, founders overpay for generic counsel and underinvest in specialized execution. Platform teams, however well-intentioned, scale poorly across sectors and introduce subtle agency problems, particularly when loyalty drifts toward the fund rather than the company. Fractional executives and market-priced specialists outperform precisely because they restore accountability and alignment.

A Series A company in Singapore does not need a generalized talent platform; it needs a domain-specific operator capable of executing under regulatory and cultural constraints. Dubai’s ecosystem has learned this through repetition: execution imported on demand outperforms standing advisory structures.

Luxembourg’s comparative advantage, by contrast, lies not in building founders, but in engineering liquidity, governance, and distribution, i.e, functions that can be priced, modularized, and exported. Unbundling restores clarity: capital allocates risk, markets supply expertise, and founders regain autonomy without illusion.

Finally, capital engineering requires distance. The most counterintuitive lesson of small ecosystems is psychological. Proximity erodes judgment faster where social graphs are dense. The healthiest investor–founder relationships are not intimate, but legible. Distance depersonalizes allocation decisions, transforming them from social acts into structural ones. Quantitative reserve management, explicit follow-on criteria, and disciplined concentration are not expressions of coldness; they are mechanisms that prevent favoritism, sunk-cost bias, and reputational inertia. In environments where venture capital is implicitly tied to national ambition, this discipline becomes political economy. Funds that allocate based on narrative rather than probability do not merely lose money; they misallocate national optionality. Large markets absorb such errors. Small systems remember them.

We should collectively recognize that venture capital, stripped of the mythology of the last decade, must again resemble infrastructure: quiet, disciplined, and designed for longevity. In large markets, venture capital can afford to be misunderstood. In Singapore, Dubai, and Luxembourg, it cannot. There, capital must be partially engineered (i.e., liquidity planned, advice unbundled, distance preserved) because the system itself depends on it. In small (eco)systems, venture capital is not a lifestyle industry. It is a structural one.

Few countries have invested as systematically in science and technology as Singapore. Over the last thirty years, the city-state has built and heavily supported an elaborate architecture of research institutes, translational centers, public–private laboratories, and state-backed venture programs. The aim has always gone beyond economic diversification. It was about positioning Singapore as a key node in the emerging knowledge economy: a place where frontier technologies are developed, commercialized, and governed — an ambition underwritten by more than SGD 60 billion in public R&D investment since 2006.

This project has yielded undeniable successes, from attracting globally respected researchers to hosting pharmaceutical manufacturing, semiconductor supply-chain nodes, and sophisticated biomedical research infrastructure. Yet the last few years — under the shadow of a global venture capital contraction — have exposed the limits of this approach. The 2024 and 2025 e-Conomy SEA reports point to a slowdown in early-stage venture formation, while PitchBook and Preqin data show multi-year declines in APAC deep-tech fundraising since 2022. In practice, most early-stage deep-tech ventures still struggle to reach meaningful commercial scale. Venture capital firms find it increasingly difficult to raise new capital without leaning heavily on government anchors. Several international players, including biotech-focused funds such as Lightstone and accelerators like Entrepreneur First, have exited Singapore, and few new deep-tech investors or company-builders have emerged in the last five years. And despite the billions invested across successive Research, Innovation and Enterprise (RIE) cycles, the conversion of research outputs into globally competitive firms remains uneven.

What happens when a developmental state attempts to engineer a deep-tech ecosystem under conditions of small-market scale, culturally embedded risk aversion, and a global investment climate that rewards contrarian boldness over administrative excellence? Deep tech, after all, cannot be steered through administrative optimization alone; it depends on contrarian judgment, tolerance for ambiguity, and strategic bets whose payoff may emerge only one or two decades later.

Singapore’s answer, I would argue, requires rethinking not only its instruments of support, but also the intellectual and cultural foundations of its innovation strategy. The country has repeatedly adapted and localised global models to its own context. Building on this tradition, Singapore could now benefit from a deeper commitment to contrarian ideas, policy experimentation, and a more distinctive “national DNA” in deep-tech development — one that does not simply import foreign playbooks, but reshapes them to the realities of a small, open, and strategically constrained economy.

The evolution of Singapore’s innovation system

Singapore’s innovation strategy is still best understood through the lens of the developmental state: long-term planning, a strong public bureaucracy, and targeted investment in priority sectors. This model has proved adaptable. The agencies that once focused on electronics, petrochemicals, and logistics now direct their efforts toward biomedical sciences, quantum technologies, advanced manufacturing, sustainable materials, and AI.

Yet this continuity obscures a significant shift. The original developmental state emerged in environments where governments could discipline domestic firms, shape capital allocation through bank-based systems, and nurture national champions in sectors with clear industrial boundaries. Deep tech does not operate under such conditions. By “deep tech,” we mean ventures grounded in scientific breakthroughs — biomedical platforms, quantum technologies, advanced materials, AI, robotics — where development is capital-intensive, slow, and dependent on specialised talent and infrastructure. Knowledge flows are global, regulatory landscapes are volatile, and competitive advantage now depends less on capex than on learning speed, network effects, scientific interpretation, and global commercial reach.

Singapore’s early deep-tech strategy — visible in the creation of Biopolis, Fusionopolis, A*STAR’s research institutes, and talent-recruitment initiatives — was analytically coherent: build research excellence first, then develop commercialization pathways on the go. Through the 2000s, the country succeeded in attracting leading scientists and building facilities that rival global innovation hubs. Yet commercialization remained modest. Much output from public research institutions translated into publications rather than venture formation, a pattern common in research-intensive systems.

The 2010s marked a shift toward entrepreneurship. A handful of startups — Structo, Horizon Quantum Computing, RWDC — demonstrated the potential for research translation and put Singapore on the deep tech map. Government-backed matching schemes and accelerators expanded the pool of founders. Corporate labs and translational units emerged in partnership with industry, creating new opportunities in medtech, advanced materials, and automation. But despite this institutional expansion, a structural bottleneck persisted: deep-tech startups struggled to scale beyond Singapore and the region, and late-stage capital remained scarce. Zimplistic, maker of the Rotimatic smart kitchen robot, is often cited as a cautionary example. Despite strong demand and more than US$50 million in funding, the company struggled with manufacturing scale, product reliability, and global distribution before ultimately being liquidated. Its trajectory illustrates Singapore’s “home-market handicap”: hardware and deep-tech companies must commercialize globally from day one, without the buffer of a large domestic market to refine manufacturing, iterate with early customers, or stabilize unit economics.

The post-2020 period has intensified the deep-tech agenda, particularly in areas aligned with geopolitical and sustainability priorities. One sees this in efforts to attract cell-therapy manufacturing, build medtech platforms, and develop quantum and photonics capabilities. Singapore has no shortage of novel technologies in the pipeline. What remains elusive is the consistent emergence of firms capable of competing globally at scale.

Structural constraints: scale, culture, and capital

These constraints run deeper than funding levels or program design.

The first is market size. Deep-tech ventures often require extensive testbeds, long regulatory pathways, supply-chain integration, and anchor customers willing to validate early proof-of-concepts. Singapore’s domestic market is too small to generate sufficient early demand for medtech diagnostics, robotics systems, or sustainable-manufacturing innovations. While government agencies sometimes act as early adopters, this cannot replicate the industrial fabric of larger economies. As a result, most successful Singapore-based deep-tech companies (e.g., BiOptic, RWDC) ultimately depended on foreign commercialization environments — an expression of the same home-market handicap noted earlier.

The second is risk culture. Singapore has invested heavily in entrepreneurial education, from university incubators to the NUS Overseas Colleges. But survey data — including the Global Entrepreneurship Monitor and studies by EDB and the Institute of Policy Studies — shows that concerns about financial risk, career disruption, and social mobility remain comparatively high. Deep tech is defined by uncertainty, long timelines, and the possibility of scientific failure. This creates a mismatch between national ambition and individual incentives. Many talented young people still gravitate toward civil service, consulting, or finance rather than high-risk research ventures. The result is a scarcity of founders willing to spend 8–10 years navigating regulation, complex science, and global market development.

The third constraint is capital structure. Seed and Series A funding are relatively well supported through state-backed co-investment schemes, angel networks, and early-stage funds. But many observers note that local deep-tech VCs are still building the capabilities required to underwrite long-cycle ventures. Others have shifted deployment abroad, attracted by denser talent pools or lower perceived risk. Very few local funds have the mandate or LP support to lead substantial Series B or C rounds; Singapore typically sees fewer than five such rounds per year, almost always led by Temasek, global corporates, or overseas investors. As a result, the availability of early institutional capital for genuinely frontier research remains thinner than headline figures suggest.

The limits of an ROI-driven approach to R&D&I

One of the least discussed but most important issues in Singapore’s innovation governance is the increasing emphasis on quantifiable returns: measurable outputs, milestone completion, near-term commercialization metrics. These tools are essential in a system where public funds must be justified. But deep tech does not conform neatly to such logic.

Research translation is inherently uncertain. Many of the breakthroughs that matter — CRISPR, mRNA therapeutics, lithium-ion batteries, photonic chips — emerged from environments that tolerated ambiguity and invested long before commercial potential crystallized. Systems that impose premature discipline risk selecting for incremental improvements rather than transformative innovation.

There is also an unintended behavioural effect. Interviews suggest that some Singapore-based founders calibrated projects around public funding schemes more than global customer needs. Compliance and milestone reporting become central activities. Internationalization is delayed until programs allow it. When public support is generous yet administratively rigid, companies can survive locally while failing to confront global markets.

This is why contrarian thinking matters. An innovation system cannot be managed purely as a portfolio of financial assets. It requires strategic bets on emerging domains that may mature only after a decade, and which conventional ROI heuristics would discourage. Without intellectual elasticity, public policy risks becoming an exercise in administrative optimisation rather than a creative act of nation-building.

Toward a distinct Singaporean deep-tech identity

If Singapore’s first twenty years of innovation were defined by institution-building and the adoption of global best practices, the next twenty demand distinctiveness. The question is not simply how to attract more researchers or fund more startups, but how to craft an innovation identity aligned with Singapore’s unique strengths: geopolitical neutrality, regulatory sophistication, financial depth, and pragmatic governance.

One possibility is to position Singapore as a regional orchestrator rather than a domestic incubator. Many Southeast Asian markets possess the scale, natural resources, and industrial challenges that require deep-tech solutions — agritech in Vietnam, medtech in Indonesia, green materials in Malaysia, and sustainable manufacturing across the region. Singapore can serve as the command center for capital formation, governance, and IP protection for these technologies, even if early deployments occur abroad. Performance Rotors, a Singapore-founded robotics company conducting industrial inspection deployments across Indonesia and Malaysia while keeping its R&D, IP, and governance anchored in Singapore, exemplifies this orchestrator model.

Another avenue is deepening translational partnerships with global industry. Singapore’s strengths in biologics, semiconductors, and specialty chemicals create opportunities for commercial testbeds rather than purely experimental labs. Companies such as Illumina, GSK, and Micron show that Singapore can host sophisticated industrial activity. The challenge is ensuring that startups — not only multinationals — are integrated into these value chains. This may require new regulatory models, procurement frameworks, or shared-infrastructure platforms that reduce barriers for smaller firms.

A related issue is talent accessibility. For deep-tech ventures to scale, Singapore must remain a place where specialised scientific and engineering talent can enter quickly and affordably. Today, the cost and administrative complexity of bringing in such profiles — especially early-career researchers and engineers — can be prohibitive for startups, given rising EP salary thresholds and the administrative burden of COMPASS evaluations, which often disadvantage junior hires who do not yet meet experience or salary benchmarks. Easing this bottleneck, without compromising standards, would materially enhance Singapore’s attractiveness as a deep-tech hub.

Most crucially, Singapore must cultivate a leadership ethos within its innovation agencies that values contrarian ideas, tolerates uncertainty, and resists emulating foreign models. The world’s successful deep-tech ecosystems — from Boston’s biotech cluster to Israel’s defence-driven innovation and Denmark’s life-science specialization — did not emerge from copying others. They articulated contextually grounded identities. Singapore must decide what its own contribution to global deep tech will be: quantum governance, sustainability technologies for tropical megacities, advanced neurodiagnostics, next-generation manufacturing, or something yet undefined.

A system optimized for administrative excellence will produce competent incrementalism. A system that allows space for contrarian thinkers — scientists, founders, policymakers — will produce transformative breakthroughs.

Leadership beyond efficiency?

Singapore remains one of the world’s most capable innovation states. Its achievements in building a deep-tech ecosystem from almost nothing are remarkable. Yet success has created its own expectations and weaknesses. The global venture-capital winter did not cause Singapore’s challenges; it illuminated them.

For Singapore to thrive in the next era of deep-tech competition, it must resist treating innovation as an efficiency exercise. The deepest questions are not administrative but philosophical: What risks is Singapore willing to take? What form of national identity should its innovation system express? And what kind of contrarian leadership will champion ideas that defy the neat logic of spreadsheets and ROI models?

Deep tech rewards courage, patience, and clarity of purpose. Singapore has these qualities in latent form; the task now is to bring them to the center of its innovation strategy. The country must move beyond adaptation and articulate a model that is not only effective but original: a deep-tech identity shaped by Singapore’s own complexity and ambition. The opportunity is simple: to build a deep-tech identity only Singapore could create.

Originally published on Asia Tech Lens on November 26, 2025.

Then came the pandemic. What should have been a pause became an acceleration. Lockdowns digitised everyday life, drawing a decade of adoption into two years. The flood of quantitative easing in the United States and Europe poured into global venture markets, inflating valuations from Singapore to Jakarta. Funds that had once hesitated to cross the Pacific now deployed aggressively, chasing category leaders at multiples once reserved for Silicon Valley.

Between 2020 and 2022, Southeast Asia experienced its own “golden years”: Grab’s listing on Nasdaq, Sea Group’s extraordinary rise, and record-breaking rounds for startups from logistics to fintech.

By mid-2023, the cycle turned. Inflation returned, rates rose, and liquidity receded as quickly as it had appeared. The region entered its first true venture winter, one not defined by crisis, but by sobriety. The exuberance of the previous years had built infrastructure, talent, and legitimacy; it had also revealed how fragile the underlying economics remained.

Funding slowed, valuations corrected, and many “regional champions” found themselves over-extended. Yet what looks like a contraction is actually a reordering: capital, talent, and ambition are rediscovering proportion. The structure of venture capital in Asia is being rewritten not by crisis, but by clarity.

The capital cycle reveals who was swimming naked

Every downturn separates those who live off fees from those who live off performance. Large LPs are redrawing their maps worldwide. In the first phase of this correction, most retreated toward the safety of scale, i.e., channelling capital into a few global megafunds whose brand and infrastructure offered predictability. That phase has largely played out. More recently, a second movement has begun: capital is trickling back toward smaller, early-stage vehicles whose alignment, cost discipline, and hunger for alpha now stand out against an industry still digesting its excesses.

Between these two poles lies a fragile middle. Funds in the US$200–500 million range (large enough to collect fees but too small to influence outcomes) face the hardest reckoning. In Southeast Asia, where most exits still fall below US$100 million and IPO windows remain closed, their economics simply do not work. At Tin Men Capital’s recent annual gathering in Singapore, several managers acknowledged the new reality: survival will depend on specialisation, not scale.

A new allocation logic is taking shape

The next decade will not reverse this polarisation; it will deepen it. Globally, megafunds will continue to expand, raising ever-larger pools and shaping markets from afar. Their exposure to Southeast Asia will remain selective, focused on a handful of Singapore-based or cross-border companies that fit within global growth theses. This capital will continue to anchor late-stage rounds and provide occasional liquidity, but it will not define the region’s venture fabric.

At the other end of the spectrum, small, agile funds are finding new room to manoeuvre. They can experiment with fee structures, syndicate models, or sharply verticalised theses; they can operate close to founders and local markets in ways global capital cannot. For them, scarcity is not a weakness but a discipline. They are the ones living off the carry, not the management fee, rather builders of conviction rather than distributors of capital.

Between these poles lies a narrow corridor where regional mid-sized funds must either reinvent their model or fade. The question is not one of size, but of incentive: who is structurally motivated to create real performance rather than perpetuate fundraising?

Southeast Asia’s constraints have become its strengths

If capital is bifurcating, the region’s structure helps explain why. Unlike the United States or China, Southeast Asia does not offer a seamless, billion-dollar domestic market. It is a mosaic of ten economies, each with distinct regulations, consumers, and currencies. That balkanisation discourages pure-scale strategies but rewards depth of expertise.

Sector knowledge now compounds faster than capital. Funds that specialise (in climate tech, logistics, healthcare, or fintech infrastructure) find proprietary deal flow where generalists see noise. Frontier technologies like AI or robotics may struggle to find exits beyond a handful of global acquirers, but sector-focused companies enjoy multiple paths: regional corporates, family conglomerates, or trade buyers.

Meanwhile, the quality of founders has improved. Corporate layoffs and global realignments have pushed seasoned operators to build their own ventures. Talent once lost to the Bay Area is returning to Singapore, Kuala Lumpur, and Bangkok. With fewer speculative investors, the ecosystem feels smaller, but also more serious.

The invisible hand of recalibration

Sovereign giants like Temasek and GIC illustrate the shift. Both have scaled back direct early-stage exposure, opting instead for fund-of-funds and co-investments, often outside the region. Their retreat leaves room for a new generation of local GPs, funds that can read the market in its own dialects and operate at founder speed.

The story, then, is not of retreat but of realignment. Capital is concentrating where it is most efficient, while creativity migrates to the edges. The next wave of Southeast Asian tech will emerge from this tension: between the institutions that industrialised venture and the craftsmen now rediscovering it.

Toward an age of proportion

Booms celebrate scale; winters reward proportion. Southeast Asia’s next cycle will be defined by the interplay of three forces: capital, reorganised and polarised; craft, rediscovered by smaller funds and founders; and constraint, the structural feature that forces both to become smarter.

If the last decade was about building the region’s digital foundations, the next will be about refining them: less rush, more resilience. In the long run, that may prove to be Southeast Asia’s most valuable innovation of all.

Photo by CJ Dayrit on Unsplash

Too often, something is called “venture building” out of convenience rather than conviction. In theory, the model promises to make innovation less risky by creating startups with better access to capital, talent, and distribution. In reality, many players emphasize the building aspect — staffing teams, producing slide decks, and launching barely proven minimum viable products — without anchoring their efforts in a coherent and validated venture thesis. This has been prevalent among the first generation of venture builders, that have been driven by corporations, agencies, and consultancies. With this group, process often replaces conviction, and volume of ventures stands in for substance. Often, the result is companies scale before they should.

There has to be a better way for Southeast Asia to do venture building.

What “building” really means

If venture is the bet, building is the work. But what constitutes building in this region can vary dramatically.

In traditional tech startups, it could focus on distribution and user acquisition — think consumer tech like Carousell or Grab. In deeptech startups such as Eureka Robotics, building might mean structuring intellectual property partnerships with universities and navigating grant regimes. In legacy sectors, it involves orchestrating value chains, regulatory approvals, and often offline infrastructure. Agritech startups like FlyLab or the late eFishery are good examples.

Building is not a uniform process, so it must be adapted to the sector, regulatory context, and local talent base. This is why the copy-paste model of Western venture builders often fails to deliver results in Southeast Asia.

Building here means starting with venture logic rooted in local realities. Southeast Asia’s challenges are acute due to its diversity: It has 11 countries with distinct languages, currencies, regulatory regimes, and stages of institutional maturity. For instance, launching a fintech product entails navigating entirely different licensing, data, and compliance frameworks in Singapore, Indonesia, and Vietnam. Infrastructure gaps are equally pronounced: What scales easily in Malaysia’s capital Kuala Lumpur is probably unviable in rural Myanmar.

Unlike China, where centralized policy enables top-down industrial scaling, or the US, where regulatory uniformity and capital liquidity create smoother pathways to market, Southeast Asia requires hyperlocal orchestration. In this region, a venture builder’s success depends not just on capital and playbooks but also on local networks and the ability to adapt operating frameworks to each geography.

Emerging models

Three types of venture builders have taken root in Southeast Asia, each shaped by structural constraints and internal logic.

First, a new class of corporate venture builders is operating alongside large enterprises. When executed well, these builders can unlock proprietary access to distribution, data ecosystems, and expertise, offering startups a formidable advantage. As always, they strive to avoid common pitfalls, such as unclear mandates and limited founder autonomy. Currently, promising examples include SC Ventures (backed by global bank Standard Chartered) and Engie Factory (backed by the French utilities giant Engie).

Both use their parent companies’ ecosystems while operating with semi-independent roadmaps and governance structures, giving them strategic alignment and entrepreneurial agility. But time will tell whether they can consistently survive three to four-year corporate cycles without being swallowed or shut down by their parent.

The second type comprises self-run studios and accelerators that have survived the funding winter. Wright Partners, Creatella Ventures, or Origgin’s Venture represent independent studios, though most had to scale down or seek complementary opportunities elsewhere.

Those are complemented by Antler, Iterative, and SDTA, which offer founder-centric, process-driven acceleration models underpinned by dedicated venture funds. Acting as institutional co-founders, they embed rigor into ideation, validation, and go-to-market phases, while providing structured support systems such as shared services, founder communities, and pre-seed capital. Their long-term success, however, hinges on attracting top entrepreneurial talent and maintaining thematic clarity. In Southeast Asia, where talent is scarce and fragmented, the accelerator model is a high-stakes, winner-takes-most game. Only those with integrated investment arms or diversified revenue streams tend to endure

The third type is a more emergent category that includes what I call mission-aligned activators, which are generally backed by private equity players. They launch ventures in sectors with high regulatory or infrastructural complexity, like climate resilience, digital health, or industrial transformation. These builders often deliver durable outcomes by working with policymakers, research institutions, and investing in parts of the ecosystem that struggle to attract capital. Wavemaker Partners kickstarted Wavemaker Impact to target climate initiatives while Temasek-backed Xorazeroes in on deep tech.

Unlike corporate venture builders that serve a parent company’s larger goals, these players are externally focused, focusing on systemic change and a multipartner approach.

Embrace change

Venture builders still have a critical role even as Southeast Asia moves from success defined by blitzscale metrics to resilience, early profitability, and long-term relevance. They serve as institutional co-founders, capable of accelerating early traction, offering operational support, and facilitating ecosystem integration.

Photo by CHUTTERSNAP on Unsplash

Yet their contribution must also be calibrated. A venture builder is not simply an enabler: It holds equity in the startups it launches, and its involvement must evolve alongside the company.

To be truly effective, a venture builder must weigh its value over time — what its contributions are, and when and how they are recognized. Short-term formats, whether fixed-term accelerators or sprint-based incubation programs, often fall short of what deep venture building requires. Building a company is a multiyear arc, not a three-month exercise.The transition from de-risking to scaling is particularly delicate. In the early stages, the builder’s presence can be immensely valuable, helping de-risk a new technology, determine a go-to-market hypothesis, or secure early capital.

But as a startup matures, that hands-on support can shift from helpful to heavy-handed, limiting agility and founder autonomy. Without a clear plan for evolving roles and incentives, builders risk becoming a bottleneck. This has implications for governance, liquidity design, and incentive structures. Builders and founders must jointly plan how equity, cash contributions, and strategic alignment will be reevaluated.

In many cases, the builder must transition from co-pilot to board-level enabler, or even exit entirely to make room for growth capital. Navigating this change transparently is crucial not only for startup success but also for the sustainability of the venture-building model.

Builders with skin in the game

Too many venture builders in Southeast Asia focus on short-term, quantitative outputs: startups launched, cohorts completed, or demo days hosted. But counting founders or pitch decks is not the same as building durable companies. The region needs builders focused on medium-term, qualitative outcomes: those who co-own risk with founders and commit to the long haul. This means corporations must evolve to see venture building as more than a PR play. It is a long-term strategy that requires internal alignment, talent autonomy, and incentive symmetry.

Similarly, government and development actors must recognize the value of venture builders in frontier markets, particularly in climate, health, and infrastructure tech. The emerging generation of builders must be as diverse as the challenges they tackle. They must operate as orchestrators, not just operators, and be measured not by the number of startups launched, but by how many durable ventures they help create.

If we can get this right, venture building will not just be a buzzword, but a foundational lever for Southeast Asia’s inclusive growth.

Originally published at https://www.techinasia.com on July 23, 2025.

Traditional venture capital has struggled with deep tech because of its long time horizons and high risk-to-reward ratio. Many investors still hesitate to back projects that require a decade of development before seeing meaningful revenue. Meanwhile, solo founders often lack the infrastructure and strategic connections needed to scale their innovations beyond the lab.

The startup studio model

This is where startup studios provide a solution. By building companies from the ground up, startup studios reduce risk, accelerate execution and optimize the startup journey through structured venture creation, pre-established networks and hands-on support. This model has already proven its efficiency in sectors like biotech and advanced manufacturing.

A growing number of deep-tech hubs are recognizing the value of startup studios. In Southeast Asia, for example, deep-tech investment has grown, yet many startups continue to struggle with cross-border expansion and industrial integration. By aligning corporate resources, academic research and venture-building expertise, startup studios offer a structured solution to bridge gaps and help technologies reach commercialization faster.

Privileged access

The defining characteristic of successful startup studios is privileged access-unique resources, networks and opportunities that are difficult for traditional startups to obtain. In deep tech, this can shorten development timelines and reduce inefficiencies.

Another significant advantage studios offer is early access to industrial and technology partnerships. The path to commercialization in deep tech requires collaboration with tier 1 players in manufacturing, supply chains and regulatory bodies. Without these connections, startups face bottlenecks in prototyping, production scaling and certification processes. Studios pre-negotiate industry partnerships, ensuring portfolio companies can integrate into existing value chains.

Another crucial dimension of privileged access is capital efficiency and early-stage funding acceleration. Many deep-tech startups face the challenge of securing long-term capital while still in the pre-revenue phase.

Studios solve this by injecting their capital early on and streamlining the fundraising process through pre-vetted investor networks and strategic co-investors. This model enables startups to spend more time on execution and less on securing funding rounds. The ability to offer early access to corporate buyers, government-backed financing and alternative funding sources also ensures studios can provide capital beyond traditional VC constraints.

The venture studio playbook

Studios engineer ventures through an adaptive, portfolio-driven strategy. One critical advantage of the model is the ability to kill ideas early and repurpose assets. In a traditional VC-backed startup, failure often means complete shutdown, but studios own a majority stake in their portfolio companies and can quickly pivot underperforming teams into new projects. This increases capital efficiency. The biotech firm Recursion Pharmaceuticals, for example, pivoted multiple times before refining its AI-driven drug discovery model, ultimately securing major partnerships.

Cross-border expansion is another area where startup studios shine. In deep tech, scaling internationally is often a necessity due to fragmented regulatory landscapes and supply chain dependencies. Many investors view international expansion too early as a risk, but in reality, I’ve found deep-tech startups must establish a global footprint from day one. Studios mitigate the risks of premature expansion by pre-building strategic entry points in key markets.

Studios also redefine the founder-equity model. While traditional startups allocate large equity shares to individual founders early on, studios operate differently. Typically, they retain 30% to 60% ownership in their portfolio companies, with founder equity structured to align incentives over time. This ensures that startups remain capital-efficient while attracting talent through salary security, access to pre-built operational teams and lower personal financial risk for entrepreneurs.

Challenges of the startup studio model

While startup studios offer a highly structured approach, there are challenges. One drawback is the high capital intensity required to sustain R&D cycles. Deep-tech ventures often require years of investment before commercialization. This puts financial strain on studios, especially if they rely on short-term capital structures. To address this, studios can diversify their funding sources, leveraging hybrid financing models that include government grants, corporate partnerships and patient capital from specialized investors.

Another challenge is talent alignment and founder autonomy. Since studios retain a significant equity stake, some experienced entrepreneurs may feel constrained by reduced ownership and decision making power. To attract top talent, studios can offer structured incentives, including long-term equity vesting, milestone-based bonuses and the ability to co-lead spin-offs. Aligning incentives through phased equity distribution ensures both founders and the studio remain committed to the venture’s success.

Finally, scaling across regulatory barriers remains a key difficulty in deep tech. Different regions have varying approval timelines, compliance requirements and infrastructure readiness. A solution here is for studios to pre-build regulatory playbooks and partnerships with policymakers, ensuring they can anticipate compliance hurdles and accelerate approval processes in multiple jurisdictions.

What’s next for startup studios?

As governments and investors seek more effective ways to accelerate deep-tech commercialization, studios are likely to play a larger role in shaping innovation ecosystems.

One emerging trend is the rise of corporate-backed startup studios, which integrate existing industrial ecosystems to de-risk venture building. Corporate venture studios allow large industrial players to test and spin out new deep-tech solutions while leveraging their infrastructure, supply chains and customer networks.

Another trend is the evolution of hybrid funding models. Traditional VC structures are often ill-suited for deep tech, but startup studios are pioneering alternative financing approaches, combining venture capital with government-backed grants, sovereign wealth funds and industry co-investments. These hybrid models provide deeper capital reserves and allow for longer development cycles without excessive dilution.

Finally, the talent equation is changing. Many serial entrepreneurs who previously built companies from scratch are now choosing startup studios over solo ventures. Studios provide a compelling value proposition-offering founders a pre-assembled team, operational infrastructure and financial stability from day one. This shift is already evident in North America and Europe, where some entrepreneurs are now building through studios rather than traditional VC-backed pathways.

As the deep-tech revolution unfolds, startup studios are an efficient and structured way to build companies. For founders, investors and policymakers alike, I think the question is no longer if startup studios work-but how to leverage them for maximum impact.

Originally published at https://www.forbes.com on March 25, 2025.

Over the past decade, I’ve worked at the intersection of deep tech investing, corporate innovation and startup acceleration, advising and co-building technology ventures across the U.S., Europe and Southeast Asia. I’ve seen firsthand how data infrastructure and operational control are becoming essential tools for private market outperformance.

Why Roll-Ups Are Ripe For AI

Roll-ups, where investors acquire and consolidate smaller businesses in fragmented sectors, have long been a private equity favorite. They can deliver efficiencies through scale, help negotiate better contracts, centralize functions and ultimately allow the group to be sold at a premium.

But executing a roll-up is hard. It requires deep market knowledge, relentless due diligence and seamless post-deal integration. That’s where AI is rewriting the playbook. Modern AI systems can crawl thousands of databases, parse regulatory filings and analyze web content to surface ideal acquisition targets using natural language processing. Machine learning models can flag customer churn risk, uncover margin levers and benchmark operational key performance before a term sheet is signed. After closing, AI can help facilitate faster onboarding, workflow automation, supply chain optimization and digital transformation across units.

This is no longer theory. Thrive Capital, an investor behind OpenAI and Stripe, has been fundraising for Thrive Holdings, a $1 billion permanent capital vehicle to acquire and operate “everyday” businesses, including homeowner associations and accounting firms, with AI as the operational backbone. The idea is to use algorithms and automation to drive improvements in margins and service across legacy sectors. And Thrive isn’t an outlier. This playbook builds on a proven model seen in industries like dental chains and plumbing services: Standardize systems, share overhead and scale intelligently. What’s new is that the engine now runs on code.

The Acceleration Of AI In Private Equity

AI in private markets may seem recent, but the data revolution has been gaining steam for decades. In the early 2000s, quantitative hedge funds like Renaissance Technologies led the way (paywall). Private equity generally followed with caution until firms like Two Six Capital began using data science to evaluate portfolio companies. Two Six participated in more than $27 billion worth of deals using these analytics. The pace accelerated in the 2020s, with some studies indicating that firms investing in data science capabilities outperformed their peers (paywall), highlighting a link between analytics and business success.

One example is Paris-based Jolt Capital, which developed Jolt.Ninja, an AI platform that’s been in use since 2016, according to the platform’s website. It scans the web to spot under-the-radar investment opportunities in tech firms. It tracks patent filings, executive shifts, market sentiment and financial signals, which can offer an edge in sourcing and diligence, particularly in Europe’s fragmented deep tech landscape. Another example is EQT, also in Europe, which uses its internal AI engine, Motherbrain, to help source investments.

Shaping Long-Term Plays

The other major trend I’m seeing reshape private markets is the rise of permanent capital vehicles (PCVs), investment structures without fixed exit deadlines. In my view, their popularity is likely increasing thanks to their compatibility with operationally intensive strategies like AI-led roll-ups. Traditional funds must return capital in seven to 10 years, in my experience. PCVs allow firms to take the long view, reinvest gains and build durable, cash-generating businesses over decades. It’s a model tailor-made for transformations that take time, like deploying AI across dozens of acquired companies.

Sequoia Capital helped pioneer this approach (paywall) in 2021 by launching The Sequoia Fund, a structure designed to hold public stocks indefinitely. Instead of being forced to exit positions in winners after an initial public offering, Sequoia now retains long-term upside and strategic optionality. Andreessen Horowitz took a similar approach in 2023 with its a16z Perennial Venture Capital Fund.

No public tally exists for how many firms run PCVs, but I’m seeing the trend accelerating. From my observations, top-tier firms with operational muscle and AI ambitions are increasingly choosing flexible timelines over forced exits.

Capital Meets Code: A Strategic Convergence

Together, AI-powered roll-ups and permanent capital vehicles signal a structural shift in how investment firms deploy capital and build value. I believe the boundary between late-stage venture and traditional private equity is fading, as operational control becomes the new priority. Firms are hiring engineers as core team members, building proprietary tooling and behaving less like investors and more like operators. A new class of fund is emerging: They aggregate assets, standardize them with AI and create long-term cash flow engines.

To me, this means the competitive edge is increasingly found in the data stack. This isn’t a tactical update; it’s a redefinition of what happens after the deal closes.

For other firms looking to adapt to these shifts, start by embedding technical talent—such as data scientists, machine learning engineers and AI product leads—into your core deal and operations teams. Second, consider piloting internal tools that can track portfolio performance, not only financially but also operationally, and layer in data streams that surface risks and opportunities in real time. I believe those who adapt could not only see better internal rates of return but also build a compounding, self-improving edge that defines the next generation of value creation.