BlackRock is marketing $12 billion in bonds through Project Sopaipilla Holdings, an 80/20 joint venture with Meta, to fund a roughly one-gigawatt data center campus in El Paso, Texas, slated to be online in 2028. This new deal is demonstrating a significant shift in investor sentiment compared to Meta's previous "Hyperion" data center financing in Louisiana. Investors are now seeking yields north of 7% on the El Paso bonds, marking a considerable increase from the terms Meta secured in October for the Hyperion project. Some investors are reportedly demanding a risk premium of approximately 0.4 percentage points higher than the Hyperion deal.
The structure of these financing deals, involving a special purpose vehicle majority-owned by a private capital partner, remains consistent. However, the market's response has changed dramatically. UBS strategist Matthew Mish noted the rapid pace of AI-related debt accumulation, estimated at about $100 billion per quarter, suggesting a potential concern within credit markets. Unlike the Hyperion bonds, which benefited from Meta's strong credit standing, the Sopaipilla bonds are being evaluated more on their own financial merits, leading investors to demand higher compensation for the perceived risk.
This repricing indicates that lenders are starting to differentiate between AI infrastructure projects, rather than uniformly viewing them as backed by the general prestige of hyperscale tech companies. For instance, a 0.1-percentage-point increase in costs for a $12 billion deal translates to tens of millions of dollars in additional annual interest expenses. The El Paso deal, maturing in 2048, is secured by Meta's 20-year rent payments starting in 2028, with S&P assigning an A+ rating and Fitch and KBRA giving an AA- rating. While Meta would cover cost overruns exceeding 105% of the initial budget, there is no direct pledge of physical assets.
Morgan Stanley projects global AI-related debt issuance to reach about $570 billion in 2026. The shift in pricing for the El Paso bonds is a clear indicator that the terms available in October are no longer applicable in July, implying increased scrutiny for future AI debt financing. This change highlights growing investor caution regarding the increasing volume of AI-related debt following a borrowing surge by major tech companies.
