AI Buildout Repeats Three Enron Financial Tactics, but Is Not Doomed to Crash

What Ram Bala Sees in the AI Buildout's Enron Echoes

The AI investment cycle has begun to resemble Enron in three specific ways, according to Ram Bala, associate professor of AI and analytics at Santa Clara University's Leavey School of Business — but he does not conclude that a collapse is inevitable. Bala says Enron's rise involved moving debt off the balance sheet through special vehicles, using mark-to-market accounting to book future sales as immediate revenue, and using circular transactions to make customer demand look more independent than it really was.

In the AI buildout, he argues, private credit now performs the role Enron's special purpose entities once did. Private equity investors including KKR are helping finance AI infrastructure; Nvidia gets paid upfront, while borrowers and their lenders carry the default risk. Because private credit is ultimately funded by ordinary savers through instruments such as pension funds, Bala warns that any tail risk would land on households.

Bala also points to demand forecasts that are 'marked to model, not to market,' and to circular deals such as Nvidia investing in OpenAI and OpenAI then using the cash to buy Nvidia's chips. He compares that to vendor financing, which can support liquidity but becomes risky when repeated too heavily around uncertain outcomes. Michael Burry has used the same Enron parallel to warn that history is repeating, while Bala's view is that long-term AI demand may ultimately justify today's financial structures — but 'only time will tell who is right.'

Private Credit, Mark-to-Model Demand and Nvidia-OpenAI Circular Deals

Why Private Credit Is Now the Hiding Place

Bala's most concrete claim is that private credit has replaced off-balance-sheet vehicles as the place 'where risk goes to become less visible.' The logic: an AI infrastructure loan held inside a private fund is harder for outside investors to price than a bond or syndicated bank loan, even when the borrower's ability to repay depends on unproven future demand. That does not make the arrangement fraudulent, but it does mean the default risk is concentrated with lenders and pension-backed savers rather than with the supplier that already collected its revenue.

Mark-to-Model Demand Forecasts Raise the Leverage Stakes

The second Enron echo is the use of future demand projections to justify capital spending today. Bala points to Anthropic CEO Dario Amodei's comment that his company has invested less in compute than rivals because a small error in demand projections could be the difference between success and bankruptcy. From Bala's perspective, when leading executives describe demand forecasts as fragile, borrowing against the optimistic case multiplies the cost of being wrong across the AI supply chain.

Nvidia-OpenAI Style Deals Work Like Vendor Financing

The circular relationship — an AI supplier investing in a customer that then buys the supplier's chips — is not unusual in expensive-equipment industries. Bala compares it to leasing a car with financing from the manufacturer. Vendor financing can create liquidity and real value, but his warning is that it becomes dangerous when layered on top of uncertain demand and hidden debt. That is the point where today's legal structures begin to reproduce the conditions that made Enron's collapse so damaging, even if Bala stops short of predicting the same ending.

What AI Lenders, Builders and Investors Should Watch For

For lenders, builders and investors with exposure to the AI stack, the article's specific warnings point to practical checks:

• For private credit investors and pension fiduciaries: ask whether AI infrastructure loans are underwritten to contracted cash flows or to mark-to-model demand, since Bala identifies that distinction as the core leverage risk.
• For AI infrastructure borrowers: stress-test compute commitments against Anthropic CEO Dario Amodei's observation that a small demand forecast error can separate survival from bankruptcy.
• For companies in Nvidia-OpenAI-style vendor financing loops: document whether supplier investment is creating genuine independent demand, because Bala says the risk rises when such circular deals are done 'too much with uncertain outcomes.'
• For investors weighing Michael Burry's warnings: distinguish between the Enron comparison as a signal of systemic leverage and Bala's view that this boom is not a past bubble — the path depends on whether long-term AI demand materializes as projected.