CoreWeave, an AI cloud company carrying about $50 billion in debt, watched its shares tumble this week. It wasn’t alone. Semiconductor makers and hardware firms, including Dell Technologies, posted sharp declines as investors asked an old-fashioned question: how exactly will all this borrowed money turn a profit?
The stumble comes as the AI-linked bull market marks its fourth anniversary. For four years, technology shares have carried the broader market. Now the bill for the build-out is arriving, and traders are reading the fine print.
The Financing Question
Concerns about paying for AI infrastructure have spread well beyond one company. Strategists at Deutsche Bank and RBC Capital Markets separately cut their recommendations on the technology sector. On October 9, Deutsche Bank strategists said the rotation into technology had “run far and the risk-reward no longer looks favorable in the near term.”
The Financial Times reported that investors are uneasy about technology companies borrowing heavily to build data centers without a clear picture of how they’ll make money from them. Katie Martin, the FT’s markets columnist, described the AI trade of the past 12 to 18 months as driven in large part by “hype and vibes.” She said it wasn’t clear whether the pullback was a short-term wobble or the start of a more serious downturn. South Korea’s semiconductor-heavy stock market fell 8% in a single session. Selling then spread to Europe and, to some extent, the United States.
Credit ratings now factor into the debate. The FT reported that Kroll Bond Rating Agency gave Anthropic a double-B grade in June, based on documents its reporters saw. That puts the AI lab close to investment grade, though KBRA didn’t have full access to Anthropic’s financial information. It relied on a limited set of financials and an in-house credit risk score, a rarity for a rating. Bankers acting for Anthropic have lobbied for an investment-grade rating after any IPO. Michelle Chan, the FT’s US credit correspondent, noted that KBRA isn’t one of the three major rating agencies, and others might not give Anthropic as much leeway. CoreWeave, meanwhile, wants an investment-grade rating for its bonds. A senior executive said that rating could help it tap insurance-company capital.
The Earnings Test
Wall Street now turns to the scorecards. Meta Platforms, Lam Research and Texas Instruments report, along with Tesla, Netflix, Procter & Gamble and GE Aerospace. LSEG IBES forecasts third-quarter earnings growth of more than 30% for S&P 500 companies.
The spending is staggering. Goldman Sachs estimates Meta and four other AI hyperscalers will lift capital spending to about $1.2 trillion next year. “One of the most important takeaways of earnings season is whether or not the profitability of the hyperscalers can grow faster than the pace of their capex,” said Kevin Shea, senior equity analyst at BNY Wealth.
Chuck Carlson, chief executive officer at Horizon Investment Services, framed the problem in terms of price. “With the cost of capital going up, this capex blowout is going to become more expensive, which raises the bar for returns,” he said. “Meta is going to be the first up to try to provide some color around that question.” Meta, valued at about $1.9 trillion, has seen its shares rise in recent weeks after launching its Muse AI assistant.
Rates and the Broader Market
Rising Treasury yields and the possibility of further Federal Reserve rate increases weigh on stock investors. Hot inflation data this week increased bets on another hike at the Fed’s meeting at the end of October. Still, futures pricing suggests most investors expect the central bank to wait until December to raise rates for a second time this year. Martin noted that investors bought U.S. government bonds even though inflation remained above target, reading the move as a sign that tech worries had unsettled them.
So far, the damage remains contained. The S&P 500 sits about 1% below its record high set this month, and sectors that lagged during the tech climb are picking up the slack. The Philadelphia SE Semiconductor index fell 8% this month but remains up more than 60% for the year, against a 13% year-to-date gain for the S&P 500. Garrett Melson, portfolio strategist with Natixis Investment Managers Solutions, called it “very much an on-off switch type of market. When AI is working, it sucks all the oxygen out of the room for the rest of the sectors.” He added, “the narrative in the market is a little bit jittery.” UBS Global Wealth Management analysts offered a steadier view in a Thursday note: “Another strong earnings season should help equities withstand moderately higher rates.”
Why This Matters:
The AI boom increasingly runs on debt, and debt carries a price that equity enthusiasm can hide. Rates are rising, and the Fed may tighten again, just as hyperscalers plan roughly $1.2 trillion in spending. Markets are doing their job: asking whether returns justify the outlay and repricing risk accordingly. For investors, retirement savers and lenders exposed to leveraged AI firms, this week’s earnings reports matter because they’ll show whether profits can outrun the capital being poured in. Disciplined capital allocation, not hype, will decide who comes out ahead.
- artificial intelligence
- technology stocks
- corporate earnings
- ai infrastructure
- coreweave
- anthropic
- meta platforms
- investor sentiment