Microsoft, Alphabet and Amazon are preparing to report results after planning hundreds of billions of dollars in AI spending. Smaller infrastructure firms, meanwhile, face rising borrowing costs and volatile markets. Together, the three hyperscalers hold a market capitalization of $10 trillion. The question hanging over their spending spree is whether it will produce sustainable revenue—or leave the risks to companies with less access to capital.
Who Controls the AI Bet
U.S. stocks finished mostly higher Tuesday as chip shares rebounded for a second trading session. The Nasdaq Composite rose nearly 1.2% to close near 26,477.81, its best session since Oct. 2, according to FactSet data. The S&P 500 gained 0.6% to 7,722.60, while the Dow Jones Industrial Average finished flat at 52,377.57. Investors awaited a Federal Reserve policy decision and earnings from Microsoft and Alphabet, due Wednesday afternoon.
Microsoft and Alphabet were set to report third-quarter results Wednesday, with Amazon following Thursday. Their planned spending shows the scale of the bet: Microsoft listed $175 billion in 2026 and $50 billion in the first quarter of 2027; Alphabet listed $195 billion to $205 billion in 2026, without exact guidance for 2027; and Amazon listed $220 billion in 2026, also without exact guidance for 2027.
Wall Street consensus expected Azure growth of about 45%, Google Cloud revenue growth of about 100% after 82% year-over-year growth in the second quarter, and AWS growth of about 38%. Investors want clearer evidence that the spending will bring in revenue. Rising interest rates have also increased borrowing costs.
Chris Brigati, chief investment officer at SWBC, said: “While we're expecting strong results across the board, the focus is less about the actual earnings results than about whether the Capex spending for AI growth will translate into sustainable revenue growth for these companies, which remains a big question mark.”
The Debt Beneath the Boom
Neocloud companies such as CoreWeave have faced volatility in their shares and bonds. Hyperscalers, by contrast, have established revenue sources and clearer access to equity and debt markets. The divide is plain: the biggest players can finance their build-out through broader markets, while smaller firms contend with financing costs that can make returns less certain.
CNBC's Jim Cramer pointed to CoreWeave bonds that yielded as much as 15%, as well as Nebius and Iren. CoreWeave shares rebounded almost 8% Tuesday, but Cramer said: “I have learned that after every positive story about anything in the data center, there will be a plethora of negative stories about something wrong, including Coreweave.”
The Anthropic financing deal also drew scrutiny. A group of banks and Blackstone announced a $63 billion financing package for Anthropic to buy Broadcom computing hardware. Cramer said Anthropic was borrowing to purchase Broadcom's technology rather than paying from business profits: “People don't seem to grasp that [Anthropic] got the money in the Broadcom-arranged financing deal to purchase equipment from Broadcom.”
Oil prices declined Tuesday, and the 10-year Treasury yield fell below 5%, easing pressure on companies borrowing for data centers. Cramer warned those conditions could reverse. The Federal Reserve planned to announce its latest interest-rate decision Wednesday afternoon, followed by Chairman Kevin Warsh's post-meeting news conference. The midterm elections the following week and possible flare-ups in the Middle East added potential sources of volatility.
Investors Want Proof, Not Just Promises
Meta investors sold the stock after the company raised the lower end of its 2026 capital-expenditure range to $135 billion to $145 billion, despite increased usage of its agentic AI assistant, Muse. Continued large investments can draw investor criticism, while cuts could leave companies behind competitors and weigh on infrastructure and data-center businesses. The bind is clear: spend heavily, or risk losing ground.
Seagate Technology reported fiscal first-quarter profit of $1.76 billion, or $7.59 a share, for the quarter ended Oct. 2, compared with $549 million, or $2.43 a share, a year earlier. Demand for data-storage components supported the profit as AI infrastructure construction continued.
Dan Ives, founder of Yorkville Ives, wrote in an Oct. 23 client note: “Roughly $850 billion of hyperscaler AI spend this year sits against more than $2 trillion of cloud backlog growing faster than revenue, so capacity, not demand, is the constraint.” He expected capex guidance to rise, with spending backed by contracted demand rather than capacity built on speculation.
Cramer said he remained “an unabashed believer in the AI build-out,” but wasn't ready to declare the coast clear. “With the exception of Nvidia, these data center companies don't have a lot of money. So they have to borrow the money. The rates are getting so high that the return is getting uncertain. That hasn't changed at all.”
- artificial intelligence
- hyperscalers
- capital spending
- stock market
- data centers
- seagate technology
- coreweave
- earnings reports