The Price of the Future

Never before has the world known more clearly what it wants to build. Artificial intelligence companies are scouting for data centres and the electricity systems to power them. Governments want semiconductor fabs, defence factories and supply chains less exposed to coercion. The energy transition requires grids, storage and new generation. Cities need infrastructure that survives a changing climate. Europe wants strategic autonomy. Japan wants military security. The Gulf wants resilience against regional shocks. All of them are bidding for patient capital. For India, which must finance industrialisation, urbanisation and the energy transition at once, the consequences are larger than for almost anyone.

For fifteen years after the global financial crisis, this was easy to overlook. Interest rates were exceptionally low. Central banks were enormous buyers of government bonds. Japanese savings flowed abroad in search of yield. That world is disappearing. Thirty-year US Treasury yields have climbed above 5%, their highest since 2007. Japanese government bond yields are around three-decade highs. India's benchmark 10-year yield sits near 6.8% even after monetary easing. This is not merely a bond-market story. It is better understood as congestion in the global market for long-duration capital.

Consider the claims now being placed on global savings. Governments must refinance enormous stocks of existing debt while financing ageing populations and rising defence expenditure. NATO members have committed to spending 5 per cent of GDP on defence and security-related expenditure by 2035. Economic security creates a newer category of investment including duplicate semiconductor capacity, alternative mineral processing, strategic inventories and redundant supply routes. Decarbonisation requires networks on an immense scale. The International Energy Agency estimates that more than 80 million kilometres of electricity grids must be added or refurbished by 2040. This is roughly equivalent to the entire grid that exists today.

 And AI has arrived as perhaps the largest new private claimant. Amazon, Microsoft, Alphabet and Meta are expected to spend roughly $725 billion on capital expenditure in 2026. The problem is simultaneity. Global savings are not a fixed pool. Over the five-to-ten-year horizons relevant to infrastructure, the supply of investors willing to absorb duration, inflation and political risk at a given price is not infinitely elastic. The world is running short of affordable duration.

Central banks do not control the price of the future. The Federal Reserve can strongly influence the price of overnight money. It has much less control over what an investor demands to surrender capital for 30 years. A long-term yield incorporates expected short-term rates, but also inflation uncertainty, fiscal credibility, bond supply and a term premium for locking money away for decades.

The signature of congestion is already visible. Long yields have risen even as policy rates have been cut or held. A central bank can lower its policy rate by 50 basis points without making a 25-year transmission project cheaper to finance if investors simultaneously demand greater compensation for duration and sovereign risk. That changes the economics of everything with a long payback.

AI is often seen as a software revolution. Economically, it is becoming one of the largest infrastructure programmes in history. A transmission line may last 50 years. A data-centre shell might last 30 years. Its financing could run for 15 or 20 years. But the accelerator inside it may be superseded within three years. This is AI's duration paradox. Long-lived capital is being committed to support short-lived technology. The question is whether AI-generated output will grow rapidly enough to remunerate that capital despite obsolescence and a higher cost of money.

Yet AI has an important advantage. Much of its investment is financed by extraordinarily profitable companies with huge cash flows and access to equity markets. They can absorb failure and rapid depreciation more easily than a regulated utility or an emerging-market infrastructure project.

Energy transition faces the same arithmetic. Solar-module prices fell by roughly 90 per cent. But cheap equipment does not produce cheap infrastructure. A renewable project depends on technology cost, financing cost and utilisation. The first has fallen spectacularly. The second is rising. The third depends on expensive networks, storage and flexibility. The scarce asset is shifting from the solar panel to the system around the solar panel.

That shifts comparative advantage. China's state-directed system can mobilise capital on strategic timescales. Japan long supplied enormous savings and exceptionally cheap money. As Japanese yields rise, domestic assets become more attractive, and countries accustomed to Japanese financing of infrastructure abroad will have to readjust their expectations.

 When capital becomes scarce, it produces a hierarchy. Congestion does not ration capital according to social need. It rations it according to the ability to pay. First comes sovereign refinancing, because maturing debt must be rolled over. Second comes AI and other high-return corporate investment, financed by balance sheets built to absorb obsolescence. Third comes defence and economic security, funded because the political cost of not doing so has become unacceptable. Then comes everything else.

The projects most vulnerable to expensive long-term capital are those with long paybacks, regulated or uncertain revenues, substantial institutional risk and benefits the investor cannot easily capture. That includes electricity transmission in emerging economies, climate adaptation, urban drainage, water resilience and coastal protection. Congestion therefore risks rationing precisely those investments with the highest social returns and the weakest private returns.

India sits almost exactly on this fault line. The National Infrastructure Pipeline alone envisaged investment of around Rs 111 lakh crore, and the requirements of the energy transition sit on top of it. India cannot determine the global risk-free rate. If the price of duration rises internationally, India cannot decree it down. But it can determine how much additional India risk investors demand on top of it. An investor financing a 25-year Indian asset prices currency risk, contractual uncertainty, regulatory change, delayed approvals, land acquisition and dispute resolution. When capital was abundant, governments could partly compensate for these weaknesses with subsidies and viability-gap funding. When capital becomes scarce, institutional quality itself becomes a financial asset.

A project financed over 25 years at 11 per cent requires annual debt service about 17 per cent higher than the same project at 9 per cent. Two hundred basis points of institutional risk removed is therefore worth roughly a sixth of the project, delivered every year, without a rupee of subsidy. How many basis points of risk does a measure remove from the cost of investing in India? India should judge reform through this metric. Faster dispute resolution, predictable tariff regulation, bankable power-purchase agreements, stable taxation, deeper bond markets and better currency hedging all have a financial value.

India's competition for capital is no longer simply with other emerging economies. A transmission project in India is competing at the margin with Northern Virginia data centres, European defence expenditure and G7 sovereign debt. The transmission line may have the greater social return. But the data centre will capture the capital unless India makes institutional certainty and deep domestic bond markets its primary financial asset.

In a world starved of affordable duration, markets will ration capital by balance-sheet muscle rather than social necessity, funding sovereign debt, AI and rearmament first while leaving grids, adaptation and basic infrastructure stranded at the back of the queue. The investments societies need most may be the last to be financed.