Who captures the value in Singapore’s S$37 billion frontier tech bet?

THE world is projected to spend US$2.5 trillion on artificial intelligence in 2026 – essentially the size of Canada’s gross domestic product, the 11th largest economy in the world.

The five largest US hyperscalers plan to deploy about US$700 billion in capital expenditure, consuming nearly 100 per cent of their operating cash flows, compared with a 10-year average of 40 per cent.

For the first time, these companies collectively hold more debt than cash on their balance sheets.

SpaceX’s unprecedented listing that soared to a US$2.5 trillion valuation, coupled with the anticipated initial public offering race between OpenAI and Anthropic, have only proven the public markets’ insatiable appetite for this infrastructure build-out.

On the flip side, the combined revenue of pure-play AI vendors – companies that derive the majority of their revenue directly from AI products or services – will likely be less than US$100 billion, according to industry sources. That is roughly 4 per cent of the infrastructure investment supporting it.

What these numbers show is that while the infrastructure is being built at historic scale, nearly all the enterprise value remains uncaptured.

The question that matters right now is: how do we capture it?

Singapore’s government, among a handful of others in the world, is asking this question earlier than most.

The answer, I believe, is that it will be captured at the layer where the technology is being applied, by companies that most investors are currently ignoring.

Understanding frontier tech’s three layers

Frontier technology can be understood in three layers. Each builds atop the last.

In enterprise, new layers stack on top of prior layers; they do not replace them. Standalone hype-driven technologies that fail to integrate across these layers are routinely overtaken by the next cycle.

The first layer is intelligent software as a service: companies that modernise old processes and unlock value from existing systems using machine learning.

This is the most commercially mature layer and where the majority of enterprise tech budgets sit today. Gartner estimates that AI-optimised servers alone will see a 49 per cent spending increase in 2026.

Value accrues to companies solving deployment problems: cost-efficiency, data verifiability, regulatory compliance, security.

The second is what we at Tin Men call “AI inside”: companies embedding new capabilities into the enterprise.

These go beyond optimisation by utilising whole-of-organisation data to generate previously unknowable insights, unlocking new categories of enterprise software. The investment profile requires deeper technical evaluation.

The third is deep tech: companies redefining what is possible through fundamental science and engineering research, spanning next-generation AI, advanced materials, autonomous systems and computational biology.

These companies have deep technical moats but require patient capital. Global military spending, a major catalyst for deep tech, surged to US$2.7 trillion in 2025, according to the United Nations.

The fragmentation thesis

As frontier technologies mature, they do not converge towards a few dominant horizontal platforms.

They fragment into what we call “uncountable large niches”: vertical, domain-specific applications where deep integration with enterprise workflows creates defensible positions.

Menlo reported that vertical AI has already captured US$3.5 billion in 2025. This is a threefold year-on-year increase, with healthcare AI alone capturing over 40 per cent of the market, having reached 50 per cent hospital penetration in the US in under two years.

Legal, the creator economy, and government fill the next three top categories in vertical AI.

McKinsey’s The State of AI in 2025 survey found that while 88 per cent of organisations now use AI in at least one function, only 6 per cent report meaningful bottom-line impact. This is the deployment gap.

The companies that close it are the ones that understand a specific industry’s workflow, can integrate with existing systems, handle regulatory compliance for that sector, and provide local support.

They are, as often said, the plumbers and electricians of the AI era: boring, essential and very difficult to displace once embedded.

South-east Asia’s structural advantage

South-east Asia – Singapore in particular – is well positioned in this value chain for reasons grounded in structural realities.

Singapore has built what amounts to a full-stack frontier technology position, with hardware (semiconductor fabs, quantum labs), capital (institutional and venture), governance (Asean Digital Economy Framework Agreement, Monetary Authority of Singapore regulation), and talent (various research centres of excellence) in a single jurisdiction.

The S$37 billion Research, Innovation and Enterprise 2030 plan spans quantum, semiconductors, decarbonisation and precision manufacturing.

In 2026, American quantum computing company Quantinuum chose to deploy its most powerful system outside the US in Singapore. AI adoption in small and medium-sized enterprises has also trebled in a single year.

The US-China technology decoupling adds another crucial, yet less talked-about, dimension.

Both superpowers capture around 65 per cent of global AI investment, but export controls and retaliatory restrictions are fracturing the supply chains that frontier technology depends on.

This fragmentation creates a structural need for neutral jurisdictions with strong rule of law and existing tech infrastructure; Singapore is the natural anchor for this emerging third node in the global value chain.

What this means for South-east Asia venture capital

In any well-functioning ecosystem, private capital plays a meaningful part in spurring innovation.

However, the venture capital (VC) market in South-east Asia is undergoing a severe reckoning. Only four funds reached final close in 2025, down from 33 in 2023.

No first-time manager reached final close in the second half of 2025.

Total startup funding also hit its lowest point in over six years: US$5.37 billion across 461 deals. Late-stage funding surged 140 per cent while seed-stage collapsed 50 per cent.

Capital is concentrating behind managers who can demonstrate distributions, disciplined deployment and defensible sourcing.

The traditional VC power law model, which assumes a small number of massive exits will compensate for a majority of failures, has been exposed as fundamentally misaligned with the region’s exit environment.

This is what a market clearing event looks like. The investors who survive it will be better for it.

Winners must close the deployment gap

A specific and time-bound opportunity is emerging in the intersection of these dynamics: an expanding global frontier technology value chain, geopolitical fragmentation that multiplies niches, Singapore’s full-stack policy push, and a structurally resetting venture market that is clearing out undisciplined capital.

The companies that close the deployment gap will capture the value. Those built around domain knowledge, integrated workflows and regulated environments where switching costs compound.

Their product is the workflow itself, embedded next to the system of record and trusted by the operations, risk and compliance teams that cannot afford a wrong answer.

So will the investors who back them.

Those with deep, operator-level understanding of enterprise technology in South-east Asia; those who can evaluate technical moats across both AI and deep tech while underwriting to the region’s exit environment; and those whose track records demonstrate that disciplined, fundamentals-focused venture capital can deliver actual liquidity to its investors.

South-east Asia’s frontier technology value chain is a distinct market shaped by unique regulatory architectures, enterprise buying patterns, talent flows and government co-investment.

The returns will accrue to investors who understand this difference and have built their entire approach around it.

The writer is co-founder and managing partner of Tin Men Capital, a Singapore-based venture capital firm exclusively backing enterprise tech startups operating from South-east Asia. He used AI for research and editing before submitting the draft