The fourth resource: Why compute capacity is the new oil and why India and the UAE may shape the next era of economic power

Every era of economic history is ultimately shaped by a scarce resource.

The 19th century was organised around coal. The 20th century belonged to oil. The 21st century is increasingly being shaped by something less visible but potentially more consequential: compute.

Artificial intelligence is often described as a software revolution. In reality, it is a resource revolution. The binding constraint is no longer only capital, labour, or even data. It is the capacity to generate, store, process, and govern intelligence at scale.

Compute is becoming the fourth resource.

Yet the analogy is imperfect. Unlike coal, oil, or even data, compute does not exist independently of the resources that sustain it. It consumes vast quantities of energy, requires massive upfront capital, depends on scarce technical talent, and increasingly competes for physical infrastructure and materials. In many respects, compute is not a standalone resource but a force multiplier acting upon all the others.

This creates a structural paradox. The more aggressively economies invest in intelligence infrastructure, the more they intensify demand for energy, capital, and industrial inputs. In parts of the US and other advanced economies, AI-related demand is already reshaping electricity pricing, grid investment, and industrial policy. The challenge is therefore not simply to expand compute capacity, but to ensure that the economic value created exceeds the resources consumed in producing it.

Like oil before it, compute is reshaping capital flows, industrial strategy, and geopolitical influence. But unlike oil, it is not merely extracted—it is engineered, financed, and governed. That distinction will define the next phase of global competition.

A shifting geography of power

One of the most important but underappreciated shifts in the global economy is the emergence of new alignments outside traditional western and Chinese technology blocs. Among these, the evolving India–UAE relationship is particularly significant.

At first glance, the partnership appears complementary. India brings scale, engineering talent, software capability, entrepreneurship, and one of the world’s largest digital economies. The UAE brings sovereign capital, energy security, infrastructure execution capability, and global financial connectivity.

Taken together, however, the relationship is more structural than transactional. It reflects the possibility of strategic autonomy in an increasingly fragmented global economy.

This matters because the AI era is not primarily about who builds the most advanced models. It is about who controls the infrastructure of intelligence itself.

Brookings Institution senior fellow Landry Signé has argued that narrowing the global AI divide will require deeper South–South cooperation and sustained investment in digital public infrastructure. The key insight is that AI inequality is not merely about access to algorithms, but about access to compute, energy systems, data infrastructure, and governance frameworks.

India and the UAE are unusually well-positioned to operationalise this vision.

Keynesianism for the age of compute

The current transformation can be understood through a Keynesian lens.

Twentieth-century development was driven by large-scale public investment in physical infrastructure: roads, ports, railways, and power systems. These investments created demand, productivity, and long-term growth.

In the 21st century, a similar role is being played by digital public infrastructure: compute clusters, cloud systems, AI research ecosystems, digital identity systems, payment rails, and data governance frameworks.

In this sense, we are witnessing the emergence of Keynesianism for the age of compute.

But the analogy is incomplete.

Traditional Keynesian frameworks assume relatively stable systems and predictable returns on infrastructure investment. The AI economy does not offer such stability. It is defined by rapid technological change, deep uncertainty, and strong feedback loops between technology, markets, and institutions.

This is where post-Keynesian insights become relevant: expectations, institutional design, and resilience matter as much as capital expenditure.

At the same time, the system is profoundly Neo-Schumpeterian. Artificial intelligence is not incrementally improving productivity. It is reorganising entire sectors through accelerated creative destruction. Finance, healthcare, logistics, manufacturing, education, and professional services are all being restructured around intelligence systems.

The result is a new economic regime: computational capitalism, where intelligence infrastructure becomes a central source of productivity, competitiveness, and geopolitical power.

Compute as an asset class

The most important shift, however, is financial.

Compute infrastructure is increasingly behaving like a distinct asset class, with its own return dynamics, risk structure, and sensitivity to technological cycles.

Unlike traditional infrastructure, compute assets are exposed simultaneously to semiconductor supply constraints, energy price volatility, currency fluctuations, AI demand cycles, regulatory shifts, and technological obsolescence risk.

Their cash flows are therefore not purely deterministic. They are path-dependent on adoption curves, network effects, and the pace of innovation itself.

This creates a structural mismatch between conventional financing models and the underlying economics of AI infrastructure.

Fixed-rate debt is often too rigid. Pure equity is frequently too expensive. Neither captures the embedded optionality in compute assets.

This is where hybrid instruments become critical.

Convertible bonds, contingent convertibles (CoCos), and other equity-linked structures provide a more natural financing architecture. They embed optionality into capital structures, allowing risk to adjust dynamically as technological conditions evolve.

In effect, they enable the emergence of compute-linked capital markets, where infrastructure is financed not as static balance sheets, but as state-contingent claims on future intelligence demand.

For sovereign investors, this reduces exposure to interest-rate cycles and currency volatility while aligning returns with the evolution of the AI economy.

Yet a deeper question remains unresolved.

Trillions of dollars are now being deployed into AI infrastructure globally, while the distribution of future returns remains highly uncertain. History offers repeated examples of transformative technologies that generated enormous social value while producing uneven or disappointing financial returns for early capital providers. Railways, electricity, and the internet all experienced phases in which investment significantly outpaced monetisation.

The question is not whether AI will transform the global economy. It is whether the financial system underpinning it is robust enough to survive the gap between technological promise and realised profitability.

Insurance, risk, and the architecture of intelligence

Every major infrastructure revolution has created new financial markets.

Maritime trade produced marine insurance. Industrialisation enabled project finance. Globalisation expanded derivatives and structured credit markets.

The intelligence economy will do the same.

AI systems introduce new categories of systemic risk: model failure, algorithmic bias, cyber disruption, semiconductor bottlenecks, cloud concentration, and cascading infrastructure failures.

These risks are not peripheral. They are structural.

As AI becomes embedded in economic decision-making, entirely new insurance and reinsurance markets will emerge around these exposures.

This is particularly relevant for the emerging India–UAE corridor. India brings deep capabilities in quantitative modelling, software engineering, and data-driven risk analytics. The UAE is rapidly positioning itself as a global hub for insurance, reinsurance, and alternative risk transfer mechanisms, with growing sophistication in risk pricing and capital structuring.

Together, they could help shape the financial architecture of AI risk itself—developing instruments, markets, and governance frameworks capable of pricing, transferring, and managing uncertainty in intelligent systems at scale.

Sovereign capital and the socialisation of finance

During periods of economic uncertainty, John Maynard Keynes argued for the socialisation of investment as a response to the structural limitations of private capital.

The compute economy may require a modern adaptation of that insight—not the socialisation of production, but the socialisation of finance.

AI infrastructure has characteristics that private markets struggle to finance efficiently: high upfront costs, long investment horizons, uncertain revenue profiles, and significant strategic externalities.

Left entirely to market forces, investment may become both insufficient and excessively concentrated.

This is where sovereign wealth funds, development institutions, pension funds, and other long-horizon capital pools become central. Their advantage lies in their ability to absorb uncertainty over decades rather than quarters.

The India–UAE partnership offers a compelling institutional model. India’s innovation ecosystem and engineering base can be combined with the UAE’s sovereign capital and infrastructure expertise to create financing structures capable of supporting strategic compute assets beyond short-term market constraints.

In this sense, the corridor is not merely about data centres or AI models.

It is about the architecture of financing intelligence itself.

Conclusion: the fourth resource

Oil defined the 20th century because it powered industry.

Compute will define the 21st because it powers intelligence.

Yet the central question is not who builds the largest models or the most advanced infrastructure. It is who builds the institutions, capital markets, risk-transfer systems, and governance frameworks capable of sustaining the intelligence economy over decades rather than quarters.

The emerging India–UAE partnership offers one possible answer.

By combining technological capability, sovereign capital, financial innovation, and long-horizon thinking, it could help define not only the infrastructure of the compute age, but also its financial architecture.

The future of the intelligence economy will not be determined solely by algorithms.

It will be determined by the institutions that finance them, the markets that price them, and the governance frameworks that ensure their benefits are widely shared.

The fourth resource has arrived.

The nations that succeed will not simply consume it.

They will define how it is built, financed, governed, and ultimately distributed.

(The author is a mathematician-turned-tail risk hedging expert, Chief Risk Strategist and advisor to sovereign institutions, Board Member at RsRL, Chief of Risk at UIB Emirates, Chief at Stochastic Commodities, and co-theorist of the Delbaen–Majumdar Theory filtering AI bias. Views are personal.)