For all the breathless talk of artificial intelligence ushering in a new era of productivity, efficiency and wealth, one uncomfortable reality is being studiously ignored by markets: AI is not cheap, it is not frictionless, and it is certainly not disinflationary.
On the contrary, the great AI gold rush threatens to re-ignite the very inflation that central banks are desperate to declare vanquished.
Investors have spent the past year congratulating themselves on a neat narrative. Inflation, they insist, belongs to the pandemic era; interest rates are on a gentle downward glide path; and technology, as ever, will save us all. This is a comforting story — and a dangerous one.
The truth is that AI is fast becoming one of the most inflationary forces in the global economy.
Behind the glossy promises of machine learning and digital transformation lies an industrial-scale arms race. The construction of vast data centres, the scramble for advanced semiconductors, and the insatiable demand for electricity are driving costs higher across multiple sectors. These pressures are not theoretical. They are real, measurable and growing — and they cut directly against the assumption that innovation automatically suppresses prices.
The technology giants leading the charge are spending sums that would have been unthinkable a decade ago. Billions are being poured into chips, cooling systems, power contracts and real estate. Energy demand alone is soaring, precisely at a moment when Western governments have constrained supply through net-zero policies and underinvestment in baseload generation. Anyone surprised by rising electricity prices in this environment has not been paying attention.
Yet markets, intoxicated by soaring share prices, have chosen to look the other way.
The complacency is striking. Equity valuations, particularly in the United States, are once again predicated on the belief that central banks will continue to ease policy indefinitely. Bond markets have priced in a benign world of falling yields and stable prices. Even as corporate executives quietly warn of margin pressure and spiralling capital expenditure, investors persist in treating inflation as yesterday’s problem.
It is not.
The idea that AI will magically boost productivity enough to offset its own costs is an article of faith rather than an established fact. Productivity gains, where they materialise, tend to be uneven and slow. Inflation, by contrast, arrives quickly when demand overwhelms constrained supply — and the AI boom is a textbook case of that imbalance.
Central banks are now caught in a trap of their own making. Having spent years flooding markets with cheap money, they are deeply reluctant to reverse course. Any renewed tightening risks puncturing the very asset bubbles they have helped inflate. But if inflation begins to creep higher — fuelled by energy prices, chip shortages and relentless investment spending — their room for manoeuvre will shrink rapidly.
The uncomfortable possibility is that policymakers may be forced into a familiar dilemma: tolerate higher inflation, or raise rates and risk a market reckoning.
Neither option is attractive. But pretending the problem does not exist is the worst choice of all.
The warning signs are already there. Technology firms are flagging rising costs. Hardware manufacturers are bracing for renewed chip price inflation. Energy grids are under strain. And governments, far from restraining spending, are embarking on fresh fiscal adventures that will only add fuel to the fire.
This is not the benign tech-driven future investors have been sold. It is an economy overheating in selective but powerful ways — precisely the kind of inflationary environment that catches markets off guard.
History offers a sobering lesson. Every technological revolution, from railways to electricity to the internet, has been accompanied by speculative excess and painful adjustment. AI will be no different. The notion that this time is immune from old economic laws is a triumph of hope over experience.
The real risk for 2026 is not that artificial intelligence fails. It is that it succeeds — expensively, voraciously, and in a manner that pushes prices higher across the economy. Inflation, once awakened, has a habit of spreading far beyond its original source.
When markets finally acknowledge this, the adjustment is unlikely to be gentle. High-growth stocks built on cheap money assumptions are particularly vulnerable. So too are governments that have grown accustomed to servicing vast debts at artificially low rates.
The AI revolution may well transform the world. But it will not repeal gravity — economic or otherwise. Inflation is stirring again, and the longer investors choose to ignore it, the harsher the eventual reckoning is likely to be.
