Bank of America warns 200bp AI spreads may pressure stocks in 2027

Bank of America strategist Michael Hartnett warned in an Oct. 25 TheStreet report that AI-linked financing could threaten stocks in 2027 if hyperscaler bond spreads widen beyond 200 basis points.

The call lands after the S&P 500 ended Oct. 22 at 7,623.68, up 11.4% in 2026 and 2.5% below its Oct. 6 record, according to the report.

Hartnett’s concern is not that artificial-intelligence investment stops. It is that the market has priced in a smooth expansion while the debt needed for chips, data centers and power infrastructure becomes more expensive or harder to secure.

TheStreet said BofA’s October fund-manager survey found just 6% of respondents expected a hard economic landing. The same report said 90% expected either a soft landing or no landing, while technology-stock positioning was at its highest overweight since August 2020.

Credit test for AI

BofA semiconductor analyst Vivek Arya estimates spending by major cloud companies could rise to $1.2 trillion in 2027 from $860 billion in 2026, almost 40% growth, TheStreet reported, citing Biggo.

Large technology companies have already leaned on bond markets. Amazon, Alphabet, Meta Platforms and Oracle issued $194 billion in bonds through July 7, compared with $108 billion throughout 2025, TheStreet reported, citing Reuters.

Hartnett warned that spreads on hyperscaler debt moving beyond 200 basis points could force selling and create broader financial stress, according to TheStreet. A basis point is one-hundredth of a percentage point, so 200 basis points equals 2 percentage points above comparable benchmark yields.

Hyperscalers are large cloud-computing companies that run extensive data-center networks. Their borrowing matters because the AI buildout depends on heavy upfront spending before all of the revenue or productivity gains are proven.

Credit markets are already separating stronger and weaker borrowers, according to TheStreet’s account of JPMorgan’s October analysis. It said credit-default swap spreads, a market price for insuring against default, ranged from 50 to 120 basis points for stronger AI companies, about 300 basis points for some infrastructure borrowers and 1,100 basis points for a heavily leveraged borrower.

The equity risk comes through expectations. A 20% shortfall against BofA’s $1.2 trillion 2027 cloud-spending forecast would leave spending $240 billion below that forecast, even though investment would still rise 11.6% from 2026, according to TheStreet.

That distinction matters for chipmakers, data-center suppliers, contractors and power-linked infrastructure companies. Stocks can fall if spending growth slows relative to forecasts, even when absolute spending continues to increase.

Three market paths

TheStreet said Hartnett framed three broad outcomes for 2027. In the first, the AI bubble keeps expanding as companies borrow, build and promise future profits while investors continue lifting valuations.

A second outcome is a more balanced market backdrop. TheStreet cited forecasts for about 15% S&P 500 earnings growth polled by FactSet and a 40-basis-point drop in Treasury yields as support for that possibility.

The third outcome is Hartnett’s downside case: a deflationary bust in which credit becomes costlier, weaker AI borrowers struggle and delayed projects feed into lower earnings expectations. That scenario would test investors’ heavy technology exposure.

The report compared the risk to the 1873 railroad crash, when borrowing-funded infrastructure spending ran ahead of demand. TheStreet cited Federal Reserve Bank of St. Louis data showing new railroad track completed fell to 1,606 miles in 1875 from 7,439 miles in 1872, a 78.4% drop.

The analogy has limits. TheStreet said the largest technology companies generate substantial cash, unlike many speculative railroad ventures, which reduces but does not eliminate the risk from more expensive financing.

The next evidence will come from hyperscaler bond spreads, capital-spending guidance, cash flow, earnings revisions and whether equity gains broaden beyond technology leaders, according to the report.

Source: trade-press report, TheStreet, Oct. 25, 2026