Major technology corporations are increasingly insulating their balance sheets from the astronomical costs of the artificial intelligence boom by pushing infrastructure risks onto private credit markets and smaller financial entities, according to reports from LinkedIn. Major players like Microsoft, Meta, and Google have structured multi-billion-dollar deals to lease computing power or finance massive facilities without directly taking on long-term debt.
Tech Giants Shift AI Risks to Private Credit Markets
For instance, Meta secured nearly $30 billion in financing to construct a massive data center in Louisiana while avoiding direct debt obligations. The project utilizes a special purpose vehicle named Beignet Investor LLC, partnered with private credit firm Blue Owl Capital. Under the arrangement, Blue Owl is on the hook for 80 percent of the financing while Meta agreed to rent the facility through a series of four-year leases. This creative financing allows Meta to categorize the funding as an operating cost rather than corporate debt in financial filings.
Similar maneuvers have been deployed across the sector. Microsoft announced a series of deals totaling tens of billions of dollars to lease computing power, while Google committed to renting computing capacity from a smaller company to sell portions to OpenAI. According to Columbia Business School accounting professor Shivaram Rajgopal, these structures simply relocate risk rather than eliminate it. Solomon Feig, a private credit lender at Pinnacle Private Credit, noted that Meta is essentially “renting risk” by paying a premium so it does not have to borrow the capital itself.
Valuation Concerns and Collateral Uncertainties
The reliance on private credit and specialized lending has triggered deep concerns among financial experts regarding asset valuation and underwriting standards. Major financial institutions recently partnered with chipmaker Nvidia to promote its computing capacity as an investable asset class. Larry Fink, CEO of BlackRock and one of Nvidia’s financing partners, compared Nvidia’s computing power assets to the mortgage-backed securities market of the 1970s. However, the financing announcements immediately raised questions regarding circular financing and how to properly value data center equipment reported by CNBC.
Unlike long-established asset classes like residential mortgages and auto loans, computing capacity is significantly harder to underwrite. Dan Alpert, founding managing partner of Westwood Capital, highlighted the extraordinarily short-term experience markets have with the asset class. Furthermore, data centers burn through high-priced graphics processing units (GPUs) on a rapid schedule. Paul Meeks, head of technology research at Freedom Capital Markets, pointed out that these GPUs depreciate on a five- or six-year schedule, leaving investors waiting years for a final scorecard on profitability.
Market Parallels and Transparency Warnings
The structure of these deals has drawn skepticism from veteran investors who warn of historical parallels to past technology infrastructure crashes. Dan Alpert compared data center-backed loans to fiber optic cable-backed loans associated with the major telecommunications bankruptcy of Global Crossing during the dot-com boom. Famed investor Michael Burry also compared the Nvidia credit agreements to Enron, writing on Substack that the arrangements bore shades of Enron’s effort to make wholesale power an investable class to prolong market momentum.
Adding to market anxiety is a distinct lack of transparency. Many of the companies managing data centers for major tech firms are privately held, engage with large start-ups, and rely on private lenders, creating obscure webs of financial stability. Wells Fargo traders noted in a recent note that Nvidia’s agreements with financial titans including KKR, Blackstone, and Apollo function as insurance for investors unfamiliar with GPU collateralization, while leaving open the question of whether AI factory loans will eventually be repackaged into collateralized loan obligations or asset-backed securities.
Worth a look
