APRIL 27, 2026
With the Strait of Hormuz closed, a widespread inflation spike is looming – Meysam Mirzadeh/Tasnim/West Asia News Agency/Reuters
We’re on the brink of a potential global recession, warned the International Monetary Fund (IMF) earlier this month – driven by the escalating conflict in the Middle East and a related rise in worldwide energy prices.
With the Strait of Hormuz impassable, a widespread inflation spike is looming, which would seriously damage the global economy.
Along with a fifth of the world’s oil and gas supplies, this dual Iran-US blockade also prevents a third of the world’s seaborne fertilizer feedstocks from reaching global markets.
That points to lower crop yields this summer – a serious energy-price spike with a food-price surge too.
A third of global helium supplies – which is is vital when making semiconductors – pass through the strait. Those chips, comparable in strategic terms to oil, enable trillions of dollars of downstream economic activity – from manufacturing to energy structures, wireless and computer data storage and defence.
On top of all that, multiple sovereign bond crises are brewing across the G7, as Western nations struggle with high fiscal deficits, tightening demographics and big government, while failing to constrain runaway welfare spending.
Large lenders are pushing up debt service costs, piling further pressure on already cash-strapped governments. The scope for systemic meltdown, caused by highly leveraged bond investors, cannot be ignored.
But what if the major threat to the global economy were none of these things – not geopolitics, spiralling food and energy prices or potential debt market chaos?
What if the main danger were elevated equity market valuations – in particular, the sky-high prices of tech stocks and firms focused on AI?
It may indeed be that global financial stability is most threatened by the very AI revolution that so many are banking on to rescue global growth.
Why? Because AI-related stocks are now trading at increasingly ridiculous valuations. And much of the huge investments going into data centres and other tech-related infrastructure are financed by debt – debt which, in turn, rests on extremely weak foundations.
I completely accept that AI, the “fifth industrial revolution”, will profoundly impact the world, potentially supercharging the global economy.
The first three “revolutions” took 250 years – steam power and mechanisation from the 1750s, then steelmaking and electrification, followed by widespread computation by the end of the 20th century.
Industrial Revolution 4.0 was the first two decades of this century – the spread of internet connectivity and “mass datafication”, meaning that much of everything ever written is now instantly and widely accessible.
And since around 2020, the AI revolution – increasingly powerful computers harnessing that connectivity and information access – has been in full swing.
The power of AI is immense. It can optimise production processes, facilitate “predictive maintenance” and automate design processes, while improving supply-chain management and quality control.
What we’re seeing now is a multibillion-dollar race, not least between US and Chinese firms, to dominate AI – the information and the processes that will increasingly drive consumer, business and indeed government behaviour across the world.
But what’s also happening is that this AI-investment frenzy is setting up equity markets for a profoundly damaging collapse.
Tech and AI-related shares now account for an astonishing 45pc of the market capitalisation of America’s S&P 500, the world’s most important equity index. That’s up from 25pc three years ago, when ChatGPT was launched.
As money has poured into AI companies, the S&P500 Cape-Shiller index – a highly influential measure of the composite cyclically-adjusted price-earnings (P/E) ratio of leading US stocks – has been driven above 40. That is well above its long-term average of 17 and much higher than 28 before the 2008 collapse.
When P/E ratios are high, equity investors are paying a lot to own a share of a company’s earnings, suggesting stocks are overly expensive. Yet these highly elevated valuations continue, despite ongoing sluggish growth, geopolitical turmoil and the prospect of serious future inflation.
In October, Nvidia – an AI chip designer that is now a “full stack” software/hardware outfit – became the first ever company to hit a $5tn (£3.6tn) market capitalisation. The value of a single company now outstrips the annual GDP of both Germany and Japan.
The trouble is that tech and AI stocks are “priced for perfection” – the idea that everything related to the development of this latest industrial revolution will turn out just fine. But it never does.
AI data centres need vast amounts of electricity, which won’t always be available. The massive affect of this tech on jobs, society and democracy means stiff regulation may will ensue, curtailing commercial dreams.
Above all, much of the investment piling into AI is highly leveraged – driven by debt provided by unregulated private credit providers. And that debt is increasingly being collateralised using the AI chips themselves.
Many investors are depreciating these chips for five to six years because of accounting purposes in their funding models. But their effective economic life is often much shorter (two to three years) because of the relentless release of more efficient hardware.
Yes, AI is economically transformational, but so were railways in the 19th century and so was the internet revolution of the late 1990s. Both drove substantial speculative investment and both caused manias, panics and crashes as multiple companies failed. History doesn’t happen in a straight line.
The big difference is that the railway and internet infrastructure created during those booms remained in place – the valuable tracks and rolling stock and the fibre-optic cables – when the various companies went bust.
The debt financing behind those investment booms was often collateralized against those valuable assets, which held their worth, being used for years by competitors and new entrants.
This AI boom is collateralized against chips prone to rapid generational obsolescence. And that, to my mind, makes what is happening now far more dangerous than previous boom-bust cycles.
Courtesy/Source: The Telegraph
