Why AI Data Center Financing Is Raising Enron and 2008 Comparisons

Steve Eisman’s concern is that off-balance-sheet financing can make AI infrastructure debt harder to see—not make the risk disappear. In Meta’s Hyperion project, a separate vehicle raised roughly $27 billion while Meta retained a minority stake and a role as tenant. That structure echoes earlier uses of off-balance-sheet entities, but the comparison is a warning about opacity and connections between borrowers and lenders, not evidence that another financial crisis is inevitable. 2311

How off-balance-sheet financing works

A special-purpose vehicle, or SPV, is a separate entity set up for a particular project. In a common data-center arrangement, the vehicle raises financing and develops or owns the facility, while a technology company holds a stake and leases the asset. The borrowing may then sit outside the parent company’s consolidated balance sheet, depending on the structure and accounting treatment. 5

Eisman argues that this can make a company’s reported leverage look lower than its broader economic commitments suggest. His comparison to Enron and pre-2008 financing focuses on the use of separate entities to keep debt less visible; it does not establish that the current deals are fraudulent or identical to the structures that preceded past crises. 2311

What Meta’s Beignet deal shows

Meta’s Hyperion data center in Louisiana was financed in part through Beignet Investor. The vehicle issued about $27.3 billion in debt; Meta holds a 20% stake, and Blue Owl-managed funds hold 80%. Because Meta is the minority owner, the project debt is not consolidated on Meta’s balance sheet under the reported treatment. 1510

That accounting treatment does not mean Meta has no economic connection to the project. Meta can lease the facility for up to 20 years and has exposure to specified costs tied to construction delays and overruns. If project performance or financing conditions weaken, creditors could face losses, while Meta’s contractual commitments could still matter to its finances. The size of those effects depends on the project’s terms and what happens; off-balance-sheet status alone does not answer that question. 5

What the auditor’s finding does—and does not—mean

Ernst & Young identified the accounting for Meta’s data-center venture as a critical audit matter, indicating that it required significant auditor judgment. The firm nevertheless issued an opinion that Meta’s financial statements were fairly presented. The distinction is important: an accounting treatment can be accepted while still involving complex judgments that investors may want to examine closely. 9

Where the risks could spread

The Bank for International Settlements has described hyperscalers’ use of both traditional bonds and off-balance-sheet arrangements to finance infrastructure, often involving private-credit firms. It says these structures strengthen links between technology companies and non-bank investors, including private-credit funds and insurers.

Banks may also be connected through funding lines to the vehicles. The BIS has identified possible transmission channels including refinancing pressure at the vehicle level, shifts in private-credit appetite and the activation of guarantees. If a project struggles, financing stress could therefore reach investors and lenders beyond the company that plans to use the data center. These are potential channels, not a prediction that losses will occur.

The practical takeaway is to look beyond whether a project’s debt appears on a hyperscaler’s balance sheet. Investors assessing exposure need to understand who borrowed, who owns the asset, what the company has promised through leases or guarantees, and which lenders and investors are connected to the project. In that sense, Eisman’s comparison is a prompt to scrutinize the financing chain—not a verdict that history is repeating itself.