9th October 2026
For months, artificial intelligence has been one of the great attractions of the stock market. Investors have poured money into the companies building AI systems, manufacturing the computer chips they need and constructing the enormous data centres required to run them.
The promise has been extraordinary. AI could transform businesses, improve productivity and create entirely new industries. Shares in companies associated with the technology have risen sharply as investors anticipate years of growth ahead.
But there is a growing question hanging over the boom: will the profits ever justify the enormous sums being spent?
That question came into sharper focus this week as technology shares fell, oil prices climbed and borrowing costs remained high. Investors are not necessarily abandoning artificial intelligence. They are beginning to look more closely at the difference between the promise of future profits and the money companies are actually making today.
A revenue figure rattles the market
One of the immediate triggers for Thursday's market fall was a report about OpenAI, the company behind ChatGPT.
OpenAI reportedly told investors that its annualised revenue was approaching $50 billion at the end of September. That was below the roughly $68 billion to $70 billion figure that had previously circulated in media reports.
The difference needs explaining. It does not mean OpenAI suddenly lost $20 billion in sales. The higher figure apparently used a different method of counting revenue connected with business partners, making a direct comparison difficult.
Annualised revenue is also not the same as a company's annual profit. It takes the current pace of revenue and projects it over a year. It does not tell us how much money the company keeps after paying for computer power, staff, research, infrastructure and other expenses.
Nevertheless, the report unsettled investors. OpenAI is central to the AI investment story, and its growth is watched closely by companies supplying the chips, computing capacity and data-centre infrastructure on which the technology depends.
On Thursday 8 October, the Nasdaq Composite, heavily weighted towards technology companies, fell about 1.3%. The S&P 500 declined around 0.5%, while the Dow Jones Industrial Average edged higher. Shares in several AI-related companies, including major chipmakers, suffered sharper falls.
The message was clear enough: when investors begin questioning the growth expectations behind AI, the consequences can spread well beyond one company.
The extraordinary cost of building the AI industry
Artificial intelligence does not run on promises alone. It requires enormous computing power, and that means expensive specialist chips, data centres, electricity supplies, cooling systems and networks.
Technology companies are committing vast sums to this infrastructure because they expect demand for AI services to keep growing. Chipmakers and data-centre operators benefit from the spending, while cloud computing companies hope to earn recurring revenue from businesses using their systems.
But there is a risk in this model. The companies supplying the infrastructure are spending heavily now in the expectation that customers will continue paying for it in the future.
What happens if demand grows more slowly than expected? What if businesses discover that some AI applications do not save enough money or generate enough extra income to justify their cost? What if competing products drive down prices?
Data centres and computer chips are expensive assets. If the expected revenue fails to materialise, companies could find themselves carrying high costs and debts without the profits they anticipated.
This does not mean the investment will fail. AI is already being used in a growing range of business activities. The uncertainty concerns how much money can ultimately be made, by whom, and how quickly.
Revenue is not profit
This is where the distinction between a successful technology and a successful investment becomes important.
A company can have rapidly rising sales and still lose money. It may need to spend heavily on research, computing power, staff and expansion simply to keep up with competitors.
Investors may tolerate losses for a time if they believe a company will eventually dominate a market and produce substantial profits. But that willingness can change when the cost of borrowing rises or growth appears less certain.
The AI industry is also highly competitive. Companies are racing to develop better models, attract paying customers and secure access to computing capacity. That competition could create useful products for consumers, but it may also make it harder for individual companies to charge high prices indefinitely.
The central question is no longer just whether AI works. It is whether the companies investing in it can turn that technology into sustainable profits on a scale that justifies their valuations.
Oil prices are adding another problem
AI concerns are not the only reason markets have become unsettled.
Oil prices have risen amid continuing uncertainty over the conflict in the Middle East. Brent crude was trading around $103 a barrel during Friday's Asian trading, after a sharp rise in the previous session.
Higher oil prices can feed through into transport, manufacturing, food production and household energy costs. If the increase lasts, it can make inflation harder to control.
That matters to stock markets because investors had hoped that inflation would ease and interest rates could eventually fall. A renewed energy-price shock could complicate that outlook.
It also puts businesses in a difficult position. They may face higher costs at the same time as customers become more cautious about spending. If they cannot pass those costs on, their profit margins come under pressure.
For households in Britain, the potential effects are familiar: more expensive fuel, pressure on the cost of living and uncertainty about the direction of mortgage and loan rates.
Why government borrowing costs matter to investors
Another source of concern is the bond market.
When government bond yields rise, investors can earn more by lending money to governments. That can make shares less attractive, particularly shares in companies whose valuations depend on profits expected many years into the future.
Higher yields also influence borrowing costs across the economy. Governments may have to pay more to refinance debt, while businesses face more expensive finance for new projects.
That is especially relevant to the AI boom, because some of the infrastructure being built requires substantial capital and long-term commitments.
A company might have a convincing plan for growth, but the financial calculation changes if it has to pay considerably more to borrow the money needed to carry it out.
The combination of expensive AI investment, high bond yields and rising energy prices is therefore more significant than any one day's stock market fall.
Is the AI bubble beginning to burst?
It is too early to reach that conclusion.
Share prices can fall because investors take profits after a strong rise, because expectations have become too optimistic, or because new information changes how much they are willing to pay for a company.
A fall in the stock market does not automatically mean that the underlying businesses are failing. Nor does a disappointing revenue comparison necessarily mean that demand for AI has collapsed.
There is a more cautious interpretation, however. If share prices have already built in years of exceptional growth, even a highly successful company may disappoint investors if its results are merely good rather than extraordinary.
This is how overvalued markets can correct themselves. The companies may continue growing, but their shares fall because investors had expected even more.
The risks are greater if heavy borrowing, expensive infrastructure and unrealistic revenue assumptions have become widespread. If investors lose confidence in those assumptions, the effect could spread through the technology sector and into the wider market.
There is also a distinction between the future of AI and the future of every company associated with it. The technology may transform business while some investors still lose money by paying too much for the shares of particular companies.
What does this mean for Britain?
Most British households do not own large portfolios of American technology shares directly. But the effects can still reach them.
Pension funds and investment funds may hold shares in international markets. A prolonged market fall can affect the value of those investments, although the impact on an individual pension depends on how it is invested and how long the money remains invested.
There can also be wider economic effects. If companies become more cautious about investment, that could affect jobs, business spending and demand for suppliers. If borrowing costs remain high, the Government may have less room to manoeuvre on public spending.
At the same time, Britain could benefit from the productive use of AI. Businesses that use the technology effectively may reduce costs, improve services and develop new products. The long-term economic benefits could be substantial, even if the investment market experiences periods of volatility along the way.
The difficulty is that those potential benefits do not arrive automatically, and they do not guarantee that every company investing in AI will succeed.
The next test is whether the numbers add up
Over the coming weeks, investors will be watching company results, revenue growth, cash flow, borrowing and the scale of further investment commitments.
They will also be watching oil prices and government bond yields, because both influence the economic conditions in which companies operate.
The key question is whether the businesses building the AI industry can demonstrate that their growing revenues will eventually produce enough profit to justify the money being spent.
If they can, recent market falls may prove to be a temporary setback in a longer period of growth. If they cannot, investors may reassess the value of companies whose share prices have depended on very ambitious expectations.
The same applies to the wider market. A few days of falling shares do not establish that a financial crisis is approaching. But the combination of high valuations, expensive investment, rising energy prices and borrowing costs deserves attention.
Artificial intelligence may indeed change the world. That is not the same as saying every investment made in its name will be profitable.
For investors, businesses and governments, the next stage of the AI boom may be less about spectacular promises and more about a much older test of business success: can the money going out eventually be justified by the money coming in?