The next empires won't be built on oil. They'll be built on data and compute.

For three hundred thousand years, human cognition was the primary engine of progress on Earth. That is not a rhetorical flourish, it is the opening line of a paper published earlier this year by economists at MIT, and it captures something that most conversations about AI still underestimate. We are not just automating tasks. We are decoupling cognition itself from biology, and that decoupling is starting to rewrite the basic assumptions of economics.

There is a word for where this leads: cognitarianism. The idea that the primary economic driver stops being human labour and becomes cognitive agents instead, systems that can plan, reason, and execute with little human oversight. It sounds abstract until you look at what is already happening. One recent industry analysis estimated the emerging agent economy at roughly 36 billion dollars in annual value on a run-rate basis, as of mid-2026, a figure that did not meaningfully exist three years ago. That number is not a forecast. It is a measurement of value already being created by deployed AI agents today.

01What changes when the scarce resource is no longer people

Economists studying this transition have started describing it as a shift in the binding constraint on growth. For most of modern history, the thing that limited how much an economy could produce was the availability of skilled human labour. As AI systems acquire the ability to plan, act, and learn with less oversight, that constraint moves. The MIT paper puts it plainly: in an economy where agents operate with real autonomy, growth is no longer limited by the scarcity of intelligence. It is limited by something else entirely, the capacity to verify, trust, and take responsibility for what those agents produce.

The European Central Bank made a related point in a speech earlier this year, noting that AI is different from earlier general-purpose technologies like electrification or the internet because it does not just make existing processes faster. It raises the productivity of the innovation process itself. A 2024 IMF study found that almost 40 percent of global employment is exposed to AI in some form, rising to roughly 60 percent in advanced economies where cognitive-task-heavy jobs are the norm. That exposure is not evenly distributed, and it is not necessarily bad news for every worker it touches. But it is a real signal that the thing economies compete over is shifting.

02The new resource map, and why nations are already racing

If data and compute are becoming what oil once was, the geopolitical response is already visible, and it moved fast. By 2026, global spending on sovereign AI infrastructure, national compute, data governance, and domestically controlled model weights, is projected to surpass 100 billion dollars. France has committed heavily to GPUs, data centres, and energy-efficient compute. Saudi Arabia's HUMAIN programme combines national data law with large-scale compute deployment and its own Arabic-language frontier models. India is deploying an 8-exaflop national AI supercomputer, roughly nineteen times its current combined AI compute capacity, built specifically so access can extend to startups, SMEs, and research institutions across the country, not just the largest players.

Singapore is very much in this race too. In May 2026, Minister for Digital Development and Information Josephine Teo announced an updated National AI Strategy built around ten priorities, with computing infrastructure and data governance named explicitly among them. The government has committed over a billion dollars to public AI research and talent development through 2030, is expanding local research compute through the National Supercomputing Centre's ASPIRE 2B platform, and has stood up a National AI Council chaired by the Prime Minister to steer the whole effort. Budget 2026 alone put S$150 million into an Enterprise Compute Initiative designed to pair companies with cloud and AI providers, alongside a 400 percent tax deduction on qualifying AI spend.

None of this means Singapore, or any mid-sized nation, is chasing full technological self-sufficiency. Analysis of Singapore's own strategy has been candid about the limits: even a wealthy, efficient, well-governed city-state remains dependent on American and Chinese infrastructure for the foundational layers of the AI stack, the chips, the frontier models, the hyperscale cloud. That is simply the honest picture. What Singapore is doing instead is maximising the ground it can actually hold: sectoral data, applied research, governance standards, and the parts of the stack where a small, disciplined country can genuinely compete.

"This is not a call for a feudal contest between nations. It is a recognition that the ground beneath the next economy is shifting, and standing still is no longer an option."

03The short-term reality inside Singapore's businesses

Step back from the geopolitics and the picture at ground level is more immediate and, frankly, more unequal. Businesses that can genuinely augment their operations with AI are already pulling ahead of those that cannot, and the gap compounds every quarter it goes unaddressed. That much is not controversial. What gets less attention is why so many small businesses are stuck on the wrong side of that gap.

It usually is not appetite. Singapore's SMEs make up 99 percent of all enterprises here and employ over 70 percent of the workforce, and most of them operate without an in-house IT or technology team, on tight margins, with small teams and no spare capacity to spend months figuring out an unfamiliar system. Layer onto that the cost structure of conventional AI providers, priced per token, scaling unpredictably with usage, with no ceiling and no path to ownership, and the barrier stops being about willingness and becomes about arithmetic. A single-person business or a lean five-person operation simply cannot absorb the same per-seat AI spend that a large enterprise treats as a rounding error.

This is the part of the cognitarian shift that deserves more attention than it gets. If cognitive agents genuinely become the primary economic engine, then access to the infrastructure that produces them stops being a convenience and starts being closer to a precondition for staying in business at all. A national strategy that gets the compute and the policy right but leaves the smallest 99 percent of enterprises unable to afford entry has only solved half the problem.

04Where Trusted Hub is putting its effort

This is exactly the gap Trusted Hub is building toward closing. Trusted ActionVerse, our no-code agentic workflow platform, exists specifically so that using AI does not require a technical team, a consultant, or a six-figure implementation budget. A guided conversation, not a coding interface, is what stands between a business owner and a working automation built on their own data.

The next stage of that work is provisioning it in smaller, genuinely independent packages, so that a trade association, a five-person operation, or a single business owner can access secure, governed AI infrastructure sized and priced for what they actually need, not scaled down from an enterprise contract that was never built with them in mind. Lowering the price of admission to the cognitive economy is not a side project for us. Given where this is heading, we think it is close to the whole point.

Whatever cognitarianism ends up being called by the time it settles into the textbooks, the underlying shift is already visible in the budgets nations are committing and the businesses being left behind while they wait. Singapore's answer has been to compete hard on the parts of the stack it can actually control. Ours, closer to home, is to make sure that competition includes everyone, not just the businesses large enough to already be in the room.