The debate over whether the artificial intelligence (AI) spending boom is primed for a bust continues to rumble.
On the one hand, Anthropic and OpenAI, the ChatGPT owner, are reportedly preparing themselves for stock market flotations with price tags in the hundreds of billions – if not trillions – of dollars.
On the other, Apple is the only one of the so-called Magnificent Seven whose shares are trading at a new all-time high as 2025 draws to a close – and it seems to be spending the least of any of them on AI.
In this context, a number of market shrewdies continue to point this column toward the history books and areas where investment is modest, even below historic averages, in the view that any upside surprise in demand will meet limited supply growth.
The net result could be much stronger prices, profits and cash flow than expected for producers. In this respect, the debate over oil is no less fierce than it is over AI.
Boom and bust
Many a stock market boom, from American railroads in the 19th century to broadband and 3G mobile telecoms equipment in the late 20th, has been fuelled by substantial capital investment in the latest technology and hopes for the associated productivity benefits.
The strong performance of shares in AI-related companies feels similar – but that raises the risk that the endgame may be the same. Capacity is built too quickly, demand is slow to catch up, profits disappoint and inevitably, spending is slashed.
Previously copious cash flow is already drying up at some of the biggest AI spenders, with the result that they are turning to debt at best – or worse, more opaque means of financing their investments through an interlocking web of supply agreements.
Even Nvidia’s balance sheet is showing strain as inventory and trade receivables rise faster than sales while growth in cash flow lags the rate of increase in revenues and profit.
These may not be red flags but they certainly have a pinkish hue – and investors are paying attention. Using the S&P Global 1200 indices as a benchmark, information technology is no longer the best-performing sector of 2025 to date, ranking third behind communication services and utilities.
This is a bit deceptive in some ways.
The biggest stocks in communication services by index weighting are Alphabet, Meta Platforms and Tencent, while utilities are being boosted as investors seek out providers of the energy that powers and cools the data centres and servers needed for the computing and inference capabilities of the large language models that in turn underpin AI.
All of this brings into relief once more the lopsided nature of the US – even global – stock markets. Technology and AI-related stocks dominate, often in unexpected ways.
Meanwhile, a poor second-half showing from sectors such as financials, industrials and real estate suggests there is some lingering disquiet over the global economic outlook, especially as more defensive areas such as healthcare and energy are doing a little better.
Power play
Energy’s gentle rally is surprising in the context of weak oil and gas prices – as well as hopes for peace in both the Middle East and Ukraine, where settlements could ease worries over the supply of hydrocarbons and perhaps lessen concerns over the importance of long-term independence.
A longer perspective than six months suggests that energy stocks are still in the doldrums.
The stock market capitalisation of the S&P Global 1200 Energy sector is just 3.3pc of that of the overall S&P Global 1200 index. That is not far from 2020’s Covid and lockdown-inspired low of 2.5pc in 2020 and well below the long-term average of 7.7pc, let alone 2007’s peak of 13.8pc.
This lack of interest reflects the view that renewables will take the lead in energy provision and also the risk that governments will continue to apply lofty tax rates if oil and gas producers’ profits stay higher for longer than expected.
The oil majors are getting the message. Their restrained capital expenditure budgets are a stark contrast to those of the AI hyperscalers.
Such relatively modest spending also suggests that supply of hydrocarbons will only grow slowly and that it will take time to bring on new fields, even if demand for energy worldwide continues to increase as the globe grapples with how best to manage the transition toward renewables.
This could yet lead to more interest in hydrocarbon assets and higher commodity prices, especially if sticky inflation or ongoing debt accumulation by Western governments persuade investors to trust “hard” assets more than “paper promises” such as cash and bonds.