World's biggest tech companies are racing to dominate AI. But at what cost?

Microsoft's blockbuster results suggest AI can deliver. But as Big Tech pours hundreds of billions into chips, data centres and infrastructure, investors are asking whether the returns will ever justify the cost.The latest earnings suggest artificial intelligence is already creating enormous business value. (Representative photo)As Big Tech pours hundreds of billions of dollars into artificial intelligence, debt markets are beginning to ask a question that equity investors have largely ignored: Can companies eventually earn enough to justify the spending?For more than two years, the artificial intelligence boom has been powered by optimism.Every new AI model, every record-breaking Nvidia earnings report and every multibillion-dollar investment announcement reinforced the belief that the industry was entering a once-in-a-generation technological revolution. Investors rewarded companies that spent aggressively, convinced that today's costs would eventually translate into tomorrow's profits.That narrative, however, is beginning to face its first meaningful test. RECORD AI SPENDINGThe latest quarterly earnings from Microsoft and Meta show why the debate over AI spending is becoming increasingly nuanced.Microsoft delivered another blockbuster quarter, with revenue rising 18% to $90 billion, Azure revenue surging 43%, Microsoft Cloud revenue climbing 27% to $59.3 billion and Microsoft 365 Copilot crossing 30 million paid users. Investors rewarded the results despite the company spending $41 billion on capital expenditure during the quarter, sending its shares sharply higher in after-hours trading. Meta, however, painted a more complicated picture. Revenue climbed 28% to $60.8 billion, but free cash flow plunged 91% as spending on AI infrastructure accelerated.The company raised the lower end of its annual capital expenditure guidance to between $130 billion and $145 billion, underscoring how expensive the AI race has become. Its shares fell sharply after the results as investors questioned whether spending was running ahead of returns.Those contrasting reactions capture the divide now emerging across financial markets.Equity investors continue to reward companies that can demonstrate tangible AI-driven growth, while debt investors are becoming increasingly focused on whether the industry's unprecedented borrowing binge can eventually produce sustainable returns.DEBT CHALLENGE FOR AI COMPANIESThe change is not coming from the stock market, where AI companies still command enormous valuations. It is emerging from the far quieter world of corporate debt, where investors are demanding higher premiums to lend money to some of the companies leading the AI race.While that does not suggest these firms are in financial trouble, it indicates that bond investors are becoming more cautious about the scale and pace of AI spending.The concern stems from a simple economic reality.Building artificial intelligence is proving to be one of the most capital-intensive technology investments in modern history. Training advanced AI models requires vast data centres, thousands of high-performance chips, dedicated networking equipment and enormous amounts of electricity. Companies are spending billions of dollars long before they can be certain how much revenue these investments will ultimately generate.To finance that expansion, many are borrowing heavily.According to estimates cited by market participants, AI-related companies have raised roughly $236 billion through debt markets in just the first five months of 2026, an extraordinary pace that reflects the industry's relentless push to build infrastructure before competitors do.Borrowing is hardly unusual for large corporations. The world's biggest companies routinely issue bonds to finance acquisitions, expand operations or invest in future growth. What is unusual is the growing nervousness among the very investors who buy that debt.One way to measure that anxiety is through credit default swaps (CDS), financial contracts that function much like insurance on corporate bonds. If investors believe the risks associated with lending to a company are rising, the cost of buying that insurance increases.Imagine two people applying for a bank loan. Both have stable jobs, but one suddenly takes on several expensive mortgages at once. Even if that person continues to earn a healthy salary, the bank may begin to see the loan as carrying greater risk. The same principle applies to companies. Rising borrowing costs do not necessarily signal that a business is in distress, but they do suggest investors are becoming less comfortable with the amount of debt it is taking on.That shift is becoming increasingly visible across the AI sector.Nvidia recently recorded the largest single-day increase in the cost of insuring its five-year debt since such contracts began trading. Oracle's five-year credit default swaps widened to around 215 basis points, implying that investors would now pay about $215,000 annually to insure every $10 million of the company's debt against default.Similar moves have been observed in debt linked to Alphabet, Amazon, Meta, Broadcom and SpaceX, reflecting a broader reassessment of risk rather than concerns about any single company.The caution is understandable when viewed alongside the scale of investment now underway.Microsoft has committed tens of billions of dollars to expand AI infrastructure. Meta now expects to spend up to $145 billion this year on capital expenditure. Amazon continues to invest aggressively through Amazon Web Services, while Oracle has become a key infrastructure partner for AI companies, prompting it to ramp up borrowing to finance new data centres.Industry estimates suggest the largest technology companies could collectively spend well over $650 billion on AI-related infrastructure and capital expenditure in 2026, an amount unprecedented even by Silicon Valley standards.The challenge is no longer whether AI can generate revenue—Microsoft's latest results suggest it clearly can. The bigger question is whether those revenues can grow fast enough to justify the extraordinary pace of investment across the industry.THE BIG PROFITABILITY QUESTIONMany AI products are still in the early stages of commercialisation. Companies continue to experiment with subscription models, enterprise software, AI-powered search and digital assistants, but the long-term economics remain uncertain. Investors broadly agree that AI will transform industries; what remains far less certain is how quickly companies will recover the hundreds of billions of dollars they are currently investing.Some financial indicators are already reflecting that pressure.Meta's borrowing costs for financing its Texas data-centre project have moved closer to levels typically associated with lower-rated corporate debt. The company's latest results also highlighted the strain of heavy infrastructure spending, with free cash flow collapsing even as revenue continued to grow strongly.The ripple effects are also being felt beyond the United States. South Korean technology stocks, particularly AI memory-chip manufacturers such as SK Hynix and Samsung Electronics, have faced bouts of sharp selling as investors reassess expectations for future AI demand. Markets are increasingly distinguishing between enthusiasm for artificial intelligence as a technology and confidence in the economics underpinning its rapid expansion.This does not necessarily mean the AI boom is running out of steam.The latest earnings suggest artificial intelligence is already creating enormous business value. Microsoft's results demonstrate that AI investments can translate into faster cloud growth, stronger enterprise demand and new revenue streams. Meta's results, however, also illustrate the financial strain of trying to build AI infrastructure at unprecedented speed.History suggests that every technological revolution eventually reaches a point where investors stop asking whether the technology will change the world and begin asking whether the companies building it can earn enough money to justify the investment.That may be the transition markets are witnessing today.For much of the AI era, the defining question was who could build the most powerful models or deploy the largest number of chips. Increasingly, investors are asking something far more fundamental: which companies can prove that their AI spending is becoming a profitable business rather than simply an extraordinarily expensive technological race.- EndsPublished By: Koustav DasPublished On: Jul 30, 2026 10:54 IST