Bank of America flags AI debt risk for stocks in 2027

Bank of America strategist Michael Hartnett is warning that financing pressures could threaten the artificial intelligence stock rally in 2027, according to TheStreet. His concern is that investors focused on the technology’s promise are underestimating the borrowing required to sustain its infrastructure expansion.

The warning centers on the relationship between credit markets and corporate investment, rather than a prediction that demand for AI will disappear. More expensive or less accessible financing could interrupt construction plans and weaken the earnings assumptions supporting shares across the sector.

According to Seeking Alpha’s account carried by TradingView, Hartnett identified bond spreads above 200 basis points for major cloud infrastructure companies as a potential trigger for investor deleveraging. That threshold represents a borrowing premium of two percentage points over benchmark yields, signaling greater compensation demanded by lenders.

Investor expectations leave limited protection against that outcome, the reports suggest. Bank of America’s October fund manager survey showed 50% anticipating continued economic expansion without a landing, 40% expecting a soft landing and just 6% forecasting a hard landing, according to Investing.com’s report carried by StreetInsider.

The bank outlined three possible paths for 2027: a continuing AI bubble, a more balanced stock-market advance or a deflationary bust. These are alternative scenarios, not a declaration that a crash is inevitable. Hartnett’s argument is that confidence in favorable outcomes appears too strong relative to concern about the downside.

In the more constructive scenario, falling Treasury yields would accompany earnings growth and help gains spread beyond technology leaders. The Investing.com account said that case rests on consensus expectations for a 40-basis-point decline in Treasury yields and approximately 15% growth in S&P 500 earnings.

The adverse scenario instead involves tighter credit forcing reductions to projected infrastructure budgets. The same report cited a $1.2 trillion forecast for major cloud companies’ capital spending in 2027, illustrating how much anticipated activity depends on companies maintaining ambitious investment plans and obtaining the resources to execute them.

TheStreet framed the potential disappointment through a spending shortfall rather than an outright collapse. Investment below expectations could still increase from the previous year, yet leave chipmakers, construction businesses and other suppliers facing weaker orders than investors had assumed. That distinction makes the pace of growth important alongside its direction.

Hartnett’s historical comparison is the railroad investment boom, according to the reports. Its relevance is the possibility that a transformative technology can remain useful while those financing its expansion suffer losses. TheStreet also noted an important limitation: the largest technology companies generate substantial cash, unlike many speculative railroad ventures.

The next test will come through borrowing conditions, corporate spending plans and earnings revisions, rather than a single verdict on AI’s usefulness. TheStreet identified credit spreads, cash flow and broader participation in stock-market gains as indicators to watch. Whether financing remains available on workable terms will help determine which of the bank’s scenarios gains credibility.