Following the latest round of earnings reports, the combined market capitalization of the four largest US hyperscale cloud providers has declined by over one trillion dollars. Investor concerns regarding uncontrolled AI infrastructure spending, pressure on cash flows, and rising debt are simultaneously depressing stock prices and increasing demand for credit hedging.
Microsoft's stock price has fallen 27% from its recent high, Amazon's has dropped 21%, Meta Platforms, Inc.'s has decreased by 16%, and Alphabet's has declined 11%. The market's core question has shifted from "Is AI worth it?" to "Can capital expenditures be sustained?", with worries that excessive investment will lead to overcapacity and extended payback periods.
This sentiment has spilled over into the debt market. Debt investors are concerned that tech giants will continue to increase leverage in the race for superior AI capabilities, leading to wider credit spreads for their bonds and a corresponding increase in trading activity for single-name credit default swaps (CDS).
According to Bloomberg, single-name CDS referencing issuers like Meta Platforms, Inc. and Alphabet have become significantly more active over the past year. Currently, the notional amount of outstanding CDS contracts for Alphabet is approximately $8.95 billion, while for Meta Platforms, Inc. it is around $6.87 billion.
Against a backdrop of consistently rising capital expenditure guidance, Goldman Sachs anticipates that the combined capital expenditures of hyperscale cloud providers will approach $1.4 trillion for the period from 2025 to 2027. Morgan Stanley, meanwhile, forecasts that these providers' borrowing could reach $400 billion this year, up from $165 billion in 2025. The "trillion-dollar retreat" in equity value and the "hedging fervor" in credit markets are jointly forcing a re-evaluation of the pricing for "big tech risk."
Investors are accelerating their exit from technology stocks. The market value losses suffered by the four hyperscale cloud providers—Amazon.com, Microsoft, Alphabet, and Meta Platforms, Inc.—following their most recent quarterly earnings reports signify a crucial shift in market sentiment. Investors are reassessing whether the escalating AI expenditures of these companies will yield commensurate returns.
Data from Goldman Sachs Global Investment Research indicates that the capital expenditure of hyperscale cloud providers is projected to surge from a total of approximately $485 billion for the 2022-2024 period to nearly $1.4 trillion for 2025-2027. Within this, Microsoft's capital expenditure is forecast to jump most significantly, from $76 billion in 2024 to $376 billion for the 2025-2027 period. Amazon Web Services is expected to spend $321 billion, Alphabet $304 billion, and Meta Platforms, Inc. $279 billion.
Goldman Sachs analyst Shreeti Kapa noted that if this level is reached, the spending intensity would approach 1.4% of GDP, a level last seen at the peak of the dot-com bubble in the late 1990s, making it a rare occurrence in modern technological history.
The credit derivatives market is expanding rapidly. Concerns among debt investors are fueling rapid growth in the credit derivatives market. A year ago, single-name CDS contracts for many high-quality tech giants barely existed; now, they are among the most actively traded instruments. Bloomberg reports that CDS trading activity for Meta Platforms, Inc. and Alphabet has increased notably recently. Net of offsetting trades, the outstanding notional for contracts linked to Alphabet's debt is approximately $8.95 billion, while for Meta Platforms, Inc. it is around $6.87 billion.
Data from the DTCC shows that by the end of 2025, the number of dealers providing quotes for Alphabet CDS increased to six from just one in July of the previous year, while dealers for Amazon.com CDS grew to five from three. London-based hedge fund Altana Wealth purchased default protection on Oracle debt last year when the cost was about 50 basis points annually, meaning it cost $5,000 per year to insure $1 million of exposure. That cost has since risen to approximately 160 basis points.
Matt McQueen, head of US corporate credit, securitized products, and municipal banking at Bank of America, stated that banks underwriting debt for hyperscale cloud providers have become significant buyers of single-name CDS. He suggested that an expected three-month distribution period might extend to nine or twelve months, leading banks to hedge part of that distribution risk in the CDS market.
Cash flow pressures are driving debt financing. The fundamental reason tech giants are being forced to tap debt markets on a large scale is that internal cash flow is insufficient to support their level of AI investment. Estimates suggest that if capital expenditure reaches the $700 billion range in 2026, that figure would nearly equal the entire operating cash flow of the hyperscale cloud providers.
Analysis by Bank of America indicates that by 2026, only Microsoft's operating cash flow is projected to still cover its capital expenditures. Meta Platforms, Inc. has hinted it may shift from being "net debt neutral" to "net debt positive." Even if stock buybacks were halted entirely, the free cash flow of the other companies would be exhausted.
Bond issuance has reached record levels. Oracle issued $25 billion in bonds, attracting $129 billion in orders. Alphabet followed, increasing a planned $15 billion bond sale to $20 billion, with orders exceeding $100 billion. According to Morgan Stanley statistics, by the end of 2025, AI-related investment-grade debt accounted for 14% of the US IG market, making it the largest single thematic sector, surpassing the banking industry in size.
Market divergence and uncertain prospects persist. Although bond demand remains strong currently, divergence has emerged in the market. Some hedge funds view the demand for protection from banks and investors as a profit opportunity.
Andrew Weinberg, a portfolio manager at Saba Capital Management, noted that since most tech giants still maintain low leverage and their bond spreads are only slightly above the average for the corporate bond index, many hedge funds are willing to sell protection. He questioned where these credits would go in a tail-risk scenario, suggesting that in many cases, large companies with strong balance sheets and trillion-dollar market capitalizations would outperform the broader credit backdrop.
However, other traders believe risk is currently mispriced. Rory Sandilands, a portfolio manager at Aegon, stated that the sheer scale of potential debt indicates that the credit risk profile of these companies could face some pressure.
Alexander Morris, CEO of F/m Investments, warned that while the investment boom in artificial intelligence is attracting many buyers, the upside is limited and the margin for error is minimal, adding that no asset class is immune to devaluation.
Analysis from Goldman Sachs points out that to maintain the return rates investors are accustomed to, these companies would need to generate annual profits exceeding $1 trillion. However, the current market consensus for profit in 2026 is only $4.5 trillion.
The ultimate outcome will depend on whether AI investments can replicate the profitability trajectory of cloud computing—where Amazon Web Services reached breakeven within three years and achieved a 30% operating margin within a decade. Before the result of this high-stakes gamble is clear, the reaction of the bond market may provide an early answer.