Hedge Funds Dump AI Debt as Data Center Loans Hit Market

Hedge funds have executed a dramatic retreat from artificial intelligence infrastructure debt, unloading massive positions in data center loans over recent weeks. According to Andrzej Skiba, head of BlueBay US fixed income at RBC Global Asset Management, the sector has witnessed unprecedented selling pressure as levered money exits AI-related credit markets. The capitulation comes as a $500 million portfolio of data center loans hits the auction block, signaling potential stress in the AI buildout financing ecosystem.

Massive Position Unwinding Across AI Infrastructure

Skiba characterized the hedge fund exodus as a "massive capitulation" during an interview on Bloomberg Television Thursday. The selling has been concentrated in debt tied to the artificial intelligence infrastructure buildout, with positions being collapsed at an accelerated pace. Officially launched on November 5, 2026, the rare loan auction includes a $25 million slice backing a data center site that forms part of the joint Stargate project, a collaboration between OpenAI, SoftBank Group Corp., and Oracle Corp. in New Mexico.

The timing of the portfolio sale is notable, coming just days after Democratic victories in US midterm elections raised uncertainty about the regulatory and policy environment for large-scale AI infrastructure projects. The auction represents one of the largest concentrated sales of AI-related debt instruments to hit secondary markets this year.

Broader Tech Selloff Extends to Equity Markets

The debt market retreat mirrors a parallel exodus from technology equities. Hedge funds dumped tech stocks at the fastest pace in at least a decade last month, according to data from Goldman Sachs Inc.'s prime brokerage desk. Skiba noted that selling has extended beyond pure AI plays to include media and telecommunications sector debt, driven primarily by levered investment strategies unwinding positions.

• Investment-grade tech bond spreads reached their widest levels versus the broader US market since 2003
• AI companies remain active borrowers this week despite market volatility
• Multitrillion-dollar AI infrastructure buildout expected to require substantial additional debt issuance

From Skeptic to Opportunistic Buyer

Skiba's positioning reflects the complex risk-reward calculus in AI credit markets. He previously characterized artificial intelligence as a "bust" for fixed-income investors, but the recent spread widening has shifted his outlook. As borrowing costs for investment-grade technology companies soared to historic premiums, RBC has grown more comfortable owning AI-related debt at wider valuations.

The average financing cost for technology borrowers now commands its largest premium over the rest of the US corporate bond market in more than two decades. This dramatic repricing has created entry points for long-term credit investors willing to ride out near-term volatility, even as short-term levered money exits positions.

What This Means

The hedge fund capitulation in AI infrastructure debt marks a significant inflection point for financing the artificial intelligence buildout. While the immediate selling pressure creates volatility, the repricing of risk may ultimately attract longer-duration capital to support data center and infrastructure projects. As AI companies continue to tap debt markets for multitrillion-dollar expansion plans, the shift from levered hedge fund positions to traditional fixed-income buyers could provide more stable financing foundations, though at higher costs that may impact project economics and deployment timelines.