Chips from Another Jacket

A viral post circulating online laid out an eye-popping balance sheet. It claimed that if Sam Bankman-Fried’s collapsed cryptocurrency empire had held onto its venture portfolio, those startup bets would be worth 206 billion dollars. That prompted a listener to ask whether Bankman-Fried was secretly one of the greatest investors of all time, or just extraordinarily lucky. That matters because recasting a convicted operator as an investing visionary blurs the line between reckless gambles and real foresight. We will trace the actual math of that portfolio, examine whose money took the hit, and see whether bull-market winners can ever prove investment skill.

The viral claim caught fire when the recovery trust for FTX, the collapsed cryptocurrency exchange, distributed over ten billion dollars to creditors. Trust administrator John J. Ray reported "continued progress returning cash to FTX's customers and creditors." With certain customer entitlement claims reaching 105 percent, commentators assumed the firm was solvent and capable of repaying creditors in full. But that recovery rests on what the court calls a petition-date claim: payouts are calculated in cash based on frozen November market-bottom prices, not what those assets would fetch today.

In reality, the portfolio was far from a triumph. Alameda Research and FTX spent about five billion dollars across hundreds of investments with incomplete documentation, a spending binge that bankruptcy CEO John J. Ray told Congress left many holdings worth only a fraction of what was paid. When customer withdrawals surged, they exposed an eight-billion-dollar hole in FTX's accounts.

To plug that shortfall, bankruptcy administrators liquidated the real holdings, recovering 1.3 billion dollars from Anthropic and 1.9 billion dollars from Solana.

Those sales brought in real money, but they were nowhere close to the viral figures—which leaves open how anyone arrived at the $170 billion Anthropic fantasy.

Untangling that viral post begins with its centerpiece: that wild Anthropic fantasy. People defending Bankman-Fried point to it as evidence of early foresight, since Alameda backed the artificial intelligence company when it was young. Yet as a startup brings in new funding, earlier investors run into cap-table dilution. In plain terms, your slice of the pie naturally shrinks whenever new shares are issued to incoming backers. The viral claim skipped that math completely, taking an early ownership share and multiplying it against massive later valuations. In the real world, court filings show the bankruptcy estate sold off the entire holding in 2024.

Next came Cursor, the AI coding tool, where the viral post credited FTX with a 3 billion dollar stake. That sounds eye-popping until you look at what actually occurred. Alameda held only a tiny pre-seed slice, and you simply cannot count something you do not own. During the restructuring, court administrators arranged a secondary buyback, handing that equity directly to the founders at their original purchase price. Counting that as a 3 billion dollar windfall is pure comedy, because any boom that happened afterward belonged to the founders who bought their startup back.

The other names on the meme collapse just as quickly. Supporters highlighted Robinhood, where Bankman-Fried had scooped up shares for 648 million dollars. Federal authorities later seized that stock, and Robinhood bought it back for about 606 million dollars. As for that supposed massive stake in SpaceX, Alameda never owned any direct equity. That exposure was strictly indirect, channeled through a separate venture firm called K5 Global.

Even researchers who calculated hypothetical paper gains without any asset sales rejected that online math. Independent figures calculated by Richard Sussman, along with analysis from The Deep Dive, indicated an unliquidated basket might have landed anywhere from 52 billion to 114 billion dollars. That confirms the headline figure was completely invented. Still, even if those bets had yielded massive theoretical wealth, it exposes a far more serious question: where did the cash to fund them actually come from?

In July 2019, Sam Bankman-Fried tweeted that Alameda's crypto account on FTX was “just like everyone else’s.” Yet according to trial testimony from co-founder Gary Wang, Bankman-Fried directed him that very day to allow Alameda to carry a negative balance, alongside an auto-liquidation exemption that let it hold losing positions indefinitely while ordinary users were margin-called. Wang testified special code funneled customer transaction deposits straight to Alameda, granting unlimited withdrawals and an alleged 65-billion-dollar credit line.

FTX customer deposits went directly into Alameda accounts, where Ray reported they were commingled with trading capital. At trial, accounting professor Peter Easton testified that as of June 2022, Alameda was supposed to hold 11.3 billion dollars in customer funds, but had only 2.3 billion in its bank accounts. Former engineering director Nishad Singh testified Alameda could “withdraw money that it didn’t have” and “lose money” that “belonged to customers.” When an executive gambles with other people's money, everything hinges on who carries the downside.

When Ray took over as CEO in bankruptcy, he said he had never seen such a complete failure of corporate controls and absence of trustworthy financial records in his forty-year career. There were virtually no internal controls and no corporate separateness, with multibillion-dollar balances tracked on QuickBooks. Ray contrasted it with Enron's financial engineering, concluding this was just plain old embezzlement.

That disorder was visible years before FTX existed. In early April 2018, Alameda executives confronted Bankman-Fried over unrecorded trades and chaotic accounting, offering a buyout if he resigned. A planning memo reviewed by Time showed a majority of staff viewed him as negligent and unethical. When he refused to step down, the four top managers and about half the staff quit.

In November 2022, the facade broke. Bankman-Fried tweeted that FTX had enough to cover all client holdings, while texting colleagues that same day acknowledging an 8.1-billion-dollar shortfall. Days later, Caroline Ellison, Alameda's chief executive, admitted to staff that customer funds had backed venture investments and loan repayments. Whether finding a few big winners across hundreds of unvetted bets demonstrates foresight, or just the law of large numbers, comes down to the rest of the ledger.

His defenders argue those home runs were not an accident, but sharp tech investing. They point to Anthropic, purchased when the young AI company had a valuation near 3.7 billion dollars, well ahead of the boom. On the stand, Bankman-Fried testified that he got into Solana at about 20 cents. To his camp, those calls show genuine vision.

Insolvency analysts see it differently: classic survivorship bias. That happens when you evaluate a whole portfolio based on a couple of breakout hits while forgetting hundreds of buried flops. John J. Ray told Congress that FTX books had incomplete paperwork for nearly 500 separate deals. Analyst Dave Manuel noted there were no real investment committees, and background vetting was practically nonexistent. When someone throws billions into hundreds of unvetted bets, landing a couple of winners is not genius. It is just probability.

Look past Anthropic, and the price of that spray becomes obvious. Manuel found that Alameda put 1.15 billion into a Kazakhstan bitcoin mining operation called Genesis Digital Assets. More than 550 million was paid out to co-founders to cash out their shares instead of funding day-to-day operations. On top of that, hundreds of millions were lost on bottom-tier tokens and exaggerated mining deals that went to zero.

The warning signs were clear inside the firm, too. Caroline Ellison testified about a financial tracking sheet where she warned that a market crash would cripple Alameda. Bankman-Fried simply commented, “yup & could also get worse.” Even after estimating a serious risk that the firm could not repay its loans, she testified he still pushed to pour money into new deals. Ellison admitted on the stand that she should have hedged their positions, but emphasized that Bankman-Fried chose to make every single investment. That comes back to who carries the downside—and how anyone can evaluate an investor's skill when the person placing the bets carried none of the risk.

Standard rules break down when an investor never risks their own money. Still, venture capitalist Jason Shuman calculated that holding winners like Anthropic and Solana could have reached 86 billion dollars. Sussman even argued Bankman-Fried might have been history's greatest venture capitalist without the crimes. In court filings before sentencing, his defense argued losses were zero because depositors were projected to receive their money back in full, while the firm's lawyers later announced repayments could include up to 9 percent interest. Highlighting the bankruptcy estate's objective of repaying creditors in full, his attorneys insisted higher market prices erased the harm.

That recovery was an accounting illusion. Under that petition-date rule from earlier, payouts were frozen at the November 2022 market bottom. As customer advocate Sunil Kavuri pointed out, depositors were never actually made whole, because they were locked into trough cash values while crypto rebounded. As lead prosecutor Nicolas Roos told jurors in closing arguments, Bankman-Fried stole the money knowing it was wrong, betting "he could walk his way out of it and talk his way out of it." That gamble ended in November 2023, when a jury convicted him on seven counts of fraud and conspiracy, bringing a 25-year prison sentence.

Market observers rejected using post-collapse rebounds to polish a felon's record. Commentator Josh Soffer noted that praising his returns is like saying someone who levered up 10,000 times on the S&P without blowing up was the best investor ever. Manuel emphasized that winning bets cannot undo the stolen deposits that will keep Bankman-Fried behind bars until 2048. When you ask who carries the downside, exchange depositors carried all the danger while Alameda pocketed the upside.

The answer to our listener is straightforward: Bankman-Fried was neither an investing visionary nor just lucky. What he ran was an asymmetric freeroll—a setup where a trader keeps the wealth and praise if a gamble succeeds, but forces depositors to shoulder the losses when it collapses. Regulatory allegations and trial testimony described Alameda tapping customer accounts as an unlimited credit line, running negative balances that exposed depositor money to risky margin trading. As prosecutors argued in court, Bankman-Fried knowingly took customer funds and lied to the public to finance speculative ventures. Winning big on stolen chips never made anyone an investing legend. It just made him a fraudster caught before he could leave the table.

For more on the venture bets behind this, explore Alameda Research. And follow business. We have more stories there, so you'll be ready when the next corporate shock lands. That's it for today. Thanks for listening, and we'll catch you next time on Angle.