Rising yields turn AI’s funding bill into a test for tech stocks

Higher borrowing costs and elevated Treasury yields are forcing investors to look beyond AI demand and scrutinise the debt, cash flow and timing behind the sector’s investment boom.

The latest wobble in technology shares is less about enthusiasm for artificial intelligence fading than about the cost of funding it. As Treasury yields have pushed higher, investors have become more cautious about richly valued growth stocks, especially those tied to AI infrastructure and large borrowing needs. In early October, the 10-year Treasury yield climbed to about 5.34%, a level not seen for years and one that Reuters said was close to multi-decade highs.

That shift matters because higher risk-free rates make future profits worth less in today’s terms. For companies whose valuations depend heavily on earnings that may arrive years down the line, the maths becomes less forgiving. Kitco reported on October 1 that the bond sell-off was powerful enough to weigh on Wall Street even as some software names advanced, underscoring how much pressure higher yields are placing on the broader growth trade.

The pressure sharpened on October 15, when Nasdaq futures fell more than 1% as investors digested fresh inflation data and concluded that rates may stay elevated for longer than hoped. Reuters said the sell-off deepened concerns around AI financing, with CoreWeave and Oracle among the names that saw notable stock declines and widening debt-related warning signs. In market terms, that suggested investors were no longer looking only at AI demand; they were also scrutinising balance sheets and the ability to service debt.

For AI infrastructure groups, that is a difficult combination. Data-centre builders and other so-called neocloud companies often need large amounts of external capital before revenue has fully caught up with spending. Kitco reported that the recent pullback has already prompted some firms to delay fundraisings or accept higher borrowing costs, while others have had to rethink listings altogether. The sector’s problem is not merely that capital is expensive; it is that the timing of the spending and the timing of the returns are now badly out of sync.

The market has not concluded that an AI bubble has burst, but it has clearly become more selective. Reuters reported on October 16 that technology stocks were still under pressure even as banks helped steady the broader market, with semiconductors and other AI-linked names remaining vulnerable to any sign that demand might slow. Cash-rich megacap companies still have room to fund investment from their own resources, but even they are exposed if growth expectations cool. For investors, the next earnings season is likely to be judged less on AI narratives than on concrete evidence of revenue, utilisation and cash generation.