FTX Didn't Outsmart Its Investors. They Forgot to Ask for the Audit.

Some of the most respected money on earth wired about $1.73 billion into FTX without ever seeing a financial statement an independent auditor had signed.⁶ Sequoia, Temasek, BlackRock, Tiger Global, Paradigm, SoftBank's Vision Fund 2 - seventy-two investors in all, the kind whose names on a cap table are supposed to mean someone checked.⁶ Then the company filed for bankruptcy, and the man brought in to run the wreckage swore under oath that FTX had been 'historically unable to produce accurate financial statements or a reliable cash forecast.'³ Not hard to read. Unable to produce. The investors had not been shown bad numbers. In many cases they had been shown numbers that were never audited at all - and they bought anyway.

The official story is that FTX was a fraud so sophisticated it fooled the smartest investors in the world. That gets the sophistication backwards. The fraud was crude - customer money handed to a sister trading firm with 'carte blanche,' a liability hidden in plain ledger entries.² What was sophisticated was the marketing around it. The investors weren't outwitted. They opted out of the one check that would have mattered.

The fraud was simple. The cover was a bull market.

Strip away the genius mythology and the mechanism is almost insultingly plain. Bankman-Fried gave Alameda Research, his own trading firm, the ability to run a negative balance of billions on FTX's exchange, unbacked by collateral, drawing on money that belonged to customers.² To make it hold together, he directed co-conspirators to alter FTX's computer code so Alameda could withdraw effectively unlimited cryptocurrency.⁴ When the liability had to live somewhere a lender might look, Alameda didn't bury it in a secret offshore account - it folded it in with legitimate third-party lender balances, so the hole sat camouflaged inside ordinary debt.² This is not the work of a criminal mastermind. It is the work of someone who correctly guessed that no one with the money would insist on tracing it.

Ray was not new to messes - he had cleaned up Enron. His point was that FTX wasn't hiding the absence of records behind clever forgeries. The records simply weren't there, and the venture-silo balance sheets that did exist were unaudited and produced under Bankman-Fried's control, which is why Ray told the court he had no confidence in them.³ The question worth sitting with is not how FTX hid this. It is how a roomful of fiduciaries decided they didn't need to look.

The diligence happened. The wrong diligence.

Here is the part that should sting. Sequoia did not skip diligence - it ran what its own letter to limited partners called a 'rigorous diligence process.'⁷ But read what that process leaned on: FTX's reported revenue of roughly $1 billion in 2021 and about $250 million in operating income, treated as inputs to the investment decision.⁷ Those figures were generated by the entity committing the fraud, and no independent auditor had ever signed them. The diligence was real. It was pointed at numbers the company invented about itself, instead of at the one external check - an audit - that exists precisely because companies lie about themselves. Sequoia wrote the whole position down to zero in a single day, losing roughly $210 million.⁷

Why settle for the company's own math? Because in 2021 the binding constraint on a hot deal was not whether the numbers checked out - it was whether you could get into the round before someone else did. FTX raised $900 million in a Series B at an $18 billion valuation in July 2021, then $400 million more six months later at $32 billion.⁶ A valuation that nearly doubles in half a year does not reward the investor who pauses to demand audited statements; it rewards the one who signs fast. The bull market didn't just inflate the price. It inverted the incentive to look closely, turning thoroughness into a competitive disadvantage. Demand an audit, lose the allocation, watch the markup happen without you.

The lie that was technically true

The cleverest thing Bankman-Fried did was not financial - it was semantic, and it tells you exactly how thin the deception had to be. He publicly assured everyone that customer deposits weren't being invested. Privately, he told a reporter the claim was 'factually accurate,' because the firm investing the money was Alameda, not FTX directly.⁸ The distinction was a fig leaf the size of a postage stamp, and it held - not because it was convincing, but because no one with capital at stake pressed on the seam. Caroline Ellison, his deputy at Alameda, later testified that he 'directed' her to 'commit crimes,' that Alameda 'took several billion dollars from FTX customers and used it for its own investments,' and that she knowingly sent 'dishonest' balance sheets to lenders.⁸ The whole edifice rested on the assumption that the people writing checks would accept what they were told. They did.

Isn't this just hindsight - wasn't FTX genuinely deceptive?

The honest objection is that this is too easy to say after the fact. FTX did actively deceive - it altered code, falsified records, and reassured customers and investors that funds were safe while concealing transactions, a finding the 2nd Circuit upheld in full in June 2026 as resting on evidence it called 'conservatively stated, robust.'⁵ So the investors were lied to. That much is true and it matters. But a deliberate lie is not a defense for a fiduciary; it is the entire reason audits exist. The standard tools - independent audited financials, verified proof of reserves, tracing where customer assets actually go - are designed for exactly the case where management lies, because management lying is the default risk, not an exotic one. The bankruptcy filing is blunt that none of the balance sheets were audited.³ The red flags were not subtle; they were structural and present at the table. They simply were not acted upon. 'We were deceived' is true. 'We were deceived because we declined to run the check built to detect deception' is truer.

FTX did not win because Bankman-Fried was a savant who built an unbreakable lie. He built a crude one - moved customer money, edited some code, hid a hole inside ordinary debt - and the smartest money in the world declined to look for it.²⁴ He was sentenced to 25 years and ordered to forfeit $11 billion, and the conviction now stands beyond appeal.⁴⁵ But the verdict that should outlive him is the quieter one: the fraud was easy because the discipline meant to stop it had become optional. The investors weren't outsmarted. They were out-hurried - and they handed over the only thing the fraud actually required, which was their permission not to check.