NEW YORK: The global artificial intelligence ( AI ) boom is entering a costly new phase as technology giants borrow hundreds of billions of dollars to build data centres, acquire advanced chips and expand computing infrastructure, raising a trillion-dollar question: who will ultimately pay for it?
Companies including Google, Microsoft, Amazon, Meta and Oracle are investing heavily in AI infrastructure, betting that demand for intelligent software and computing services will generate substantial returns.
However, the enormous scale of investment is raising concerns among economists and investors about whether future revenues can justify the spending.
According to Goldman Sachs figures cited by the Financial Times, investors have lent approximately $500 billion to AI-related companies so far in 2026, including around $200 billion to major cloud infrastructure providers.
Technology giants turn to borrowing
Major technology companies have traditionally relied on substantial cash reserves and operating profits to finance expansion.
But the extraordinary cost of developing AI infrastructure is driving greater reliance on corporate borrowing.
The Financial Times reported that major cloud infrastructure companies could borrow approximately $1 trillion by 2030 as they expand their global data centre networks.
The borrowing is also transforming corporate bond markets, with technology companies raising funds in multiple currencies to finance their projects.
Investors are becoming more cautious as companies repeatedly increase their projected capital expenditure, making future financing requirements difficult to assess.
Will AI generate enough revenue?
The central concern is whether AI services can generate sufficient income to recover the enormous investments being made today.
A Bain & Company study estimated that major technology infrastructure providers and other AI companies would need more than $4.2 trillion in additional revenue over five years to finance the expansion.
The consultancy warned that existing markets might not generate enough productivity gains to justify the expenditure, meaning entirely new commercial opportunities would be required.
Meanwhile, research cited by Reuters suggests global spending on data centres alone could exceed $30 trillion by 2050.
Economists caution that AI’s economic benefits may take years or even decades to materialise, while companies face immediate financing costs and repayment obligations.
Financial markets face growing pressure
Rising borrowing costs are adding another challenge to the AI investment race.
US government bond yields have climbed to levels not seen in more than two decades, increasing pressure on corporate financing and stock valuations.
The growing supply of technology company debt is also competing for investors’ funds at a time when global bond markets are experiencing volatility.
On Thursday, Asian stock markets declined as concerns about borrowing costs and major technology companies’ financing plans weighed on investor sentiment.
The risks extend beyond technology companies themselves.
ALSO READ: Elon Musk Accuses Indian Oligarchs Of Blocking Starlink Launch
Banks, bondholders and other lenders could face losses if AI projects fail to attract sufficient customers or if infrastructure becomes less valuable than expected.
Businesses purchasing AI services may also face pressure to demonstrate that the technology delivers measurable productivity improvements.
A revolution or an expensive gamble?
Despite the concerns, supporters believe artificial intelligence could transform industries ranging from healthcare and manufacturing to education and scientific research.
AI developers are betting that more capable systems will create entirely new markets and significantly improve productivity.
But economists point to previous technological revolutions, including railways and the internet, where infrastructure investment initially exceeded commercial returns.
The International Monetary Fund has also warned that delays in realising AI’s economic benefits could create financial risks, even as the technology contributes to growth and corporate earnings.
For now, technology companies are financing an extraordinary expansion on the expectation of future demand.
Whether shareholders, lenders and customers ultimately benefit will depend on how quickly AI delivers revenues that can support its enormous infrastructure costs.
