Big Short Investor Steve Eisman Warns AI Boom Shows Enron-Era Warning Signs

Steve Eisman, the investor whose successful bet against the U.S. housing market made him one of the most prominent figures associated with The Big Short, has raised a new warning about the rapidly expanding artificial intelligence industry. Eisman says some of the financing structures being used to fund massive AI infrastructure projects are bringing back practices he believed had largely disappeared after the collapse of Enron and the 2008 financial crisis. His concerns centre particularly on special purpose vehicles and other arrangements that can move significant amounts of debt away from the balance sheets of the companies ultimately exposed to the projects.

According to Eisman, the growing cost of AI infrastructure is pushing major technology companies to seek increasingly sophisticated ways of financing data centres, computing capacity and other infrastructure. Companies including Microsoft, Alphabet, Oracle and Meta have been investing enormous amounts of capital to support the expansion of AI, creating a financing environment that is substantially more capital-intensive than many of these businesses were in previous technology cycles. Eisman argues that investors should pay close attention to who ultimately carries the financial risk when these projects are funded through separate entities rather than directly through corporate balance sheets.

One example highlighted by Eisman involves Meta and a joint venture connected to a data-centre project. The structure reportedly raised about $27 billion in debt through an entity known as Beignet Investor LLC, allowing the financing to sit outside Meta’s conventional balance sheet even though the technology company remains closely connected to the project and its economic risks. Eisman sees such arrangements as particularly noteworthy because similar off-balance-sheet structures became controversial during previous financial crises. His comparison is not an accusation that today’s companies are committing accounting fraud, but rather a warning that investors should understand the liabilities and risks hidden within complex financing arrangements.

Eisman’s warning comes as concerns about the sustainability of the AI investment boom continue to grow. He has previously argued that the industry’s biggest vulnerability may not be chipmakers such as Nvidia or the major cloud companies funding infrastructure, but AI developers such as OpenAI and Anthropic, whose enormous computing requirements must eventually translate into sustainable revenues and profits. At the same time, other investors have raised concerns about whether the returns generated by AI will justify the extraordinary amount of capital being deployed.

The comparison with Enron therefore reflects a broader concern about transparency, leverage and the distribution of risk rather than a prediction that the AI industry will necessarily collapse. AI remains a rapidly expanding technology with substantial commercial applications, while companies continue to report strong demand for computing infrastructure. However, Eisman’s comments suggest that investors may need to look beyond headline revenue and earnings figures and examine how AI infrastructure is financed, where debt is recorded and which companies could ultimately absorb losses if projected returns fail to materialise.