The Enron Playbook: What Rickards Compares AI To

Jim Rickards drops one sentence in his AI Black Paper promo that stops you cold. AI companies, he says, are using “the same shady accounting practices as Enron and Lehman Brothers.”

That is a loaded comparison. Enron is not just a scandal. It is the scandal — the one that wiped out $74 billion in market cap, vaporized employee pensions, and destroyed one of the five biggest accounting firms on the planet.

If Rickards is comparing AI debt to that, you need to know what “that” actually was.

Here is the full story. No shortcuts. No spin. Just the mechanics, the collapse, and the question nobody in the AI hype machine wants to answer.

What Enron Actually Did

Enron was not a straightforward fraud. There was no single person cooking a single set of books. It was a system — a deliberate architecture of financial engineering designed to make the company look far healthier than it was.

Two mechanisms made it work: Special Purpose Entities and mark-to-market accounting. Understand those two things, and you understand Enron.

The SPV Machine

Start with the Special Purpose Entities (SPEs, also called SPVs). Enron created hundreds of them — shell companies, partnerships, and trusts that existed for one reason: to hold debt off Enron’s balance sheet.

Here is how it worked.

Enron would build an asset — a power plant, a pipeline, a broadband network — and then sell it to an SPE it had created. The SPE would borrow money to buy the asset. That borrowing showed up on the SPE’s books. Not on Enron’s.

If Enron’s total debt looked artificially low, its credit rating stayed high. Its stock stayed high. Executives cashed out. And nobody knew the truth until it was too late.

The accounting rules at the time said an SPE could stay off the parent’s balance sheet if at least 3% of its equity came from outside investors — people truly independent of Enron. That is the “3% rule.” In theory it was a gate. In practice it was a joke.

Enron failed that test repeatedly. The Chewco SPE — named after the Star Wars character, because Wall Street has a sense of humor — was capitalized with loans from Barclays that were backed by Enron cash collateral. The “outside” money was not at risk. It was Enron’s own cash dressed up in a different suit.

The Powers Committee report, Enron’s own internal investigation, found that Chewco and a web of related entities called the Raptors (Talon, Timberwolf, Bobcat — all Star Wars names) should have been consolidated into Enron’s financial statements from the start. When they finally were, in November 2001, the restatement wiped out $586 million in previously reported profit and added billions in hidden debt.

Enron used roughly 500 SPEs in total. Five hundred.

Mark-to-Market Accounting

The second mechanism was mark-to-market (MTM) accounting. The SEC approved Enron’s use of MTM in 1992. It was a departure from traditional accounting, where you book revenue when you actually earn it. Under MTM, Enron could book the net present value of a long-term contract as current income — today, right now, before a single dollar of cash ever changed hands.

Sign a 20-year contract to supply natural gas? Estimate the total profit over two decades. Discount it to present value. Book it as revenue this quarter.

The problem is obvious: those estimates were fantasy.

Take the Blockbuster deal. In July 2000, Enron signed a 20-year agreement to stream movies on demand. The technology barely worked. Pilot projects in Portland, Seattle, and Salt Lake City reached a few dozen apartments. Enron booked $110 million in estimated profits anyway.

Take the Sithe Energies contract — a 20-year natural gas supply deal. Sithe owed Enron $1.5 billion by the late 1990s. Enron’s own risk team estimated Sithe’s only asset was worth $400 million — nowhere near enough to pay. Enron did not write down the contract. It kept the profit on the books.

How did managers get away with this? Because their compensation was tied to those same fair-value estimates. The people valuing the contracts were the people getting paid based on those valuations. There is no better recipe for overstatement.

By the time the truth came out, Enron had recognized hundreds of millions in income that had never been collected — and never would be.

The Collapse

It happened fast.

October 16, 2001: Enron reports a $638 million loss for the third quarter and a $1.2 billion reduction in shareholder equity. The market blinks.

October 22: The SEC opens an inquiry into Enron’s related-party transactions.

October 24: Enron fires CFO Andrew Fastow, the architect of the SPE structures.

October 31: The SEC upgrades its inquiry to a formal investigation.

November 8: Enron restates its financial statements for the previous five years — 1997 through 2001. The company that reported massive profits now says it actually lost $586 million.

December 2: Enron files for Chapter 11 bankruptcy protection. Fourteen affiliated entities file alongside it.

The stock, which hit $90.75 per share at its peak, traded at $0.26 by bankruptcy.

Seventy-four billion dollars in market value. Gone.

Employees who had their retirement savings locked in Enron stock — the company barred them from selling during the collapse — lost everything. Arthur Andersen, Enron’s auditor, started shredding Enron documents on October 10. The shredders ran until November 8, when a federal subpoena finally arrived. Andersen was convicted of obstruction of justice in June 2002. The firm, one of the Big Five, was effectively destroyed — 28,000 U.S. jobs vaporized.

The Prosecution

The Enron Task Force spent years picking up the pieces.

Jeff Skilling, Enron’s CEO, was convicted in May 2006 on 19 counts including securities fraud, insider trading, and conspiracy. He was sentenced to 24 years and 4 months in federal prison and ordered to forfeit $45 million. He served about 12 years before his sentence was commuted in 2019.

Kenneth Lay, Enron’s founder and chairman, was convicted on all six counts — conspiracy, wire fraud, securities fraud — plus separate bank fraud charges. He died of a heart attack on July 5, 2006, before sentencing. His conviction was voided.

Andrew Fastow, the CFO who designed the SPE system, pleaded guilty in 2004 to two counts of conspiracy. He cooperated with prosecutors, surrendered nearly $30 million in cash and property, and testified against Skilling and Lay. The judge gave him 6 years — down from the 10 he had agreed to — citing his cooperation and the persecution his family endured. He served about 5 years.

The legislative response was Sarbanes-Oxley, passed in 2002. It tightened disclosure requirements, created new criminal penalties for corporate fraud, and established the Public Company Accounting Oversight Board. The 3% SPE loophole was eventually closed.

The Legitimate Side of SPVs

Here is the part Rickards’ promo does not emphasize.

Special Purpose Vehicles are not inherently fraudulent. They are used every day in legitimate securitization — mortgages bundled into mortgage-backed securities, auto loans pooled into asset-backed bonds, credit card receivables financed through master trusts. The structure isolates assets so investors can evaluate them on their own merits, independent of the originating company’s overall financial health.

That is how the modern credit system works. It is legal. It is regulated. And it is genuinely useful.

The problem is not the vehicle. The problem is the opacity. When you cannot see what is inside the SPV — when the assets, liabilities, and risk are hidden from investors — you cannot price the risk. Enron exploited that opacity deliberately. They created entities with no real outside capital, stuffed them with their own stock and toxic assets, and called them independent.

The SPV is a tool. Enron used it as a weapon.

Is the AI Comparison Fair?

This is where it gets interesting.

AI companies are using SPVs to finance data-center construction. Meta’s Hyperion campus in Louisiana, financed through a $27 billion SPV co-owned with Blue Owl Capital, is the biggest example. The debt sits off Meta’s balance sheet. The structure is real, and the scale is enormous.

That is NOT the same as what Enron did.

The intent appears different. Enron used SPVs to hide existing losses and fabricate earnings. AI companies using SPVs today appear to be doing legitimate project finance — ring-fencing a specific asset, securing long-term leases, and letting investors fund the build-out. The accounting rules have also changed. Post-Enron standards like ASC 842 and IFRS 16 pull many of these obligations back onto the balance sheet or into footnotes.

Rickards’ comparison sounds aggressive because it is. Calling the Hyperion deal “the same shady practices as Enron” is a stretch — on intent alone.

But here is where Rickards earns the comparison.

The structural risk is the same. Opacity. When billions in debt live inside SPVs that are not fully consolidated, investors cannot see the full picture. They see headline numbers that look healthier than the underlying reality. If AI revenue disappoints — and MIT research has found that 95% of corporate AI initiatives have so far produced no measurable ROI — those lease obligations do not disappear. They sit inside SPVs, waiting to be repriced.

The problem is not fraud. The problem is hidden leverage that nobody is pricing correctly.

Rickards’ point is structural, not moral. Systemic risk does not require criminal intent. It requires opacity, leverage, and a belief that good times will last forever. That is what broke Enron. That is what broke Lehman. And that is what Rickards is warning about now.

You can disagree with the Enron analogy. But you cannot dismiss the question.

This is part of the flackjacketfinance.com Reader HQ guide collection. For the full context, read the AI Black Paper promo review and the $200 billion debt fact-check.