Wall Street Grows Wary of AI Financing as Nvidia, AMD and Broadcom Face Pressure

The sell-off in artificial intelligence-related stocks is increasingly being driven by concerns about how companies finance their expansion and whether they can generate sufficient cash flow to support the cost of that investment, according to Baird technology strategist Ted Mortonson.

Investors are reassessing the financing arrangements and debt burdens supporting the artificial intelligence boom, putting pressure on semiconductor and data-center companies. [Photo: Coreweave]

“I think the street is tiring of the Bank of Nvidia, the Bank of AMD and the Bank of Broadcom,” Mortonson told Yahoo Finance senior reporter Brooke DiPalma, referring to the role major chipmakers play in financing or supporting the broader AI infrastructure ecosystem.

Mortonson said institutional investors were scrutinizing higher capital costs and financing arrangements that could create circular dependencies among companies investing in AI infrastructure.

The concerns intensified after reports that Anthropic told investors at a private meeting that its third-quarter revenue had reached $16.8 billion.

Investors have also been assessing the pace of revenue growth relative to the enormous sums being committed to computing infrastructure.

CoreWeave shares fell more than 15 percent at one point this week, declining, as well, in premarket trading, as investors questioned the company’s debt burden and future borrowing costs.

Mortonson identified CoreWeave, Nebius and Oracle as companies with particularly weak free cash flow relative to their investment needs. He said their reliance on external financing made them vulnerable to higher capital costs.

The sell-off has extended beyond infrastructure providers to chipmakers whose shares had benefited from the AI boom.

Companies with strong reported earnings have also faced selling pressure as investors reassess valuations and the sustainability of future growth.

A Gartner forecast cited in the report added to the uncertainty, suggesting that most independent neocloud providers could cease operating independently by 2030.

Such an outcome could involve acquisitions, consolidation or restructuring rather than necessarily meaning that all affected businesses would shut down.

The recent declines highlight a shift in investor attention from AI demand alone toward the financing structures, cash generation and returns that will determine whether the industry’s expansion remains sustainable.