Will Sam Bankman-Fried's Guilty Verdict Change Anything?

The predawn tableau outside the federal courthouse on Worth Street in lower Manhattan—with its shady band of hunched silhouettes drifting in and out of conspiratorial conversations—resembled scenes playing out elsewhere in the city at the same unseemly hour, before sample sales or at open-air drug markets. The mildly distracting intoxicant being meted out was the trial of Sam Bankman-Fried, who had been charged with seven counts of fraud related to his cryptocurrency exchange, FTX. On my first day at the courthouse, I arrived at about a quarter before six in the morning. Before I could take up a seat on one of the granite benches outside, I was summoned toward the entrance by a spirited yet vaguely rumpled figure who turned out to be David Yaffe-Bellany, the lead crypto reporter for the Times. Yaffe-Bellany is only a few years out of college, but the demands of the trial had given him a temporarily wizened affect; he had typically been arriving no later than about four-thirty in the morning to insure a place in the courtroom. He asked me to sign my name to a ripped piece of ruled paper, which showed that I was No. 17—barely inside the margin of the twenty-one daily spots available in the courtroom proper. He seemed a little apologetic about the improvised bureaucracy; the previous trust-based system—a regular queue—had been recently violated by another journalist, who had (perhaps inadvertently) cut the line. A centralized ledger had become necessary. A few minutes later, he pointed to a shaggy-haired new arrival and said, with resigned disapproval, “That’s the guy.”

Line infractions aside, the mood was convivial. Mainstream reporters mixed collegially with more domain-specific veterans from CoinDesk and Blockworks and with a diverse spectrum of crypto enthusiasts, including Tiffany Fong, a YouTuber who had scored one of the first post-collapse interviews with Bankman-Fried, and a large, fit, tattooed cigar-smoker who went by the name Taco. The atmosphere probably had something to do with the shared commitment to arrival before daybreak—and, as the trial went on, the relieved fixation with a headline story that wasn’t a cause for widespread fracture and despair. (Electronics are also strictly forbidden inside the courthouse, which allows a welcome reprieve from doomscrolling.) But it also reflected the unifying appeal of the Bankman-Fried saga. Among the many story lines on offer were tales about crypto and its discontents; soap-operatic plots about friendship, romance, family, and betrayal; philosophical determinations about how far the rules could be bent in the service of doing good; and the debate over Michael Lewis’s controversial assessment, “Going Infinite,” which came out last month. There was a surfeit of narrative intrigue to go around.

By now, the contours of the case are well known: Bankman-Fried, an M.I.T. physics graduate and the son of two Stanford Law professors, made a colossal amount of money in a very short interval, first as the founder of a trading firm, Alameda Research, that looked to exploit inefficiencies in the freewheeling, unsupervised crypto markets, and then as the founder of FTX, a crypto exchange that sold itself to customers and investors as something other than freewheeling and unsupervised. He allocated that wealth prodigiously—on venture investments, political campaigns, and marketing deals—in anticipation of its further exponential growth, all, he claimed, with an eye toward the long-term security and prosperity of humanity. All of a sudden, in November of last year, the money was gone. The missing funds included at least eight billion dollars in FTX customer assets, which he had promised to cordon off from all the other money he had sloshing around. According to the prosecution, the whole thing was more or less a shell game from the beginning: he lost a lot of money, used stolen customer funds to cover the losses, then lost those, too. Bankman-Fried has claimed that he never swindled anybody, arguing that, although he made “a number of small mistakes and a number of larger mistakes,” they’d resulted from the good-faith failures of a network of companies whose growth outstripped the development of proper corporate controls or risk management.

Stories such as Bankman-Fried’s are exercises in genre. The plot of financial fraud is almost always the same—there are minor problems that might reasonably be cleared up; a little dishonesty is in the meantime brought to bear on the problems; the problems, compounded by the dishonesty, expand beyond the point of repair—and what differentiates one from the next are subtle variations on the theme. Elizabeth Holmes was a young woman and a Stanford dropout; she had a creepy older boyfriend, liked to cosplay as Steve Jobs, and duped Henry Kissinger. Sam Bankman-Fried is a nerd and the son of principled intellectuals; he had a co-conspiring nerd for a girlfriend, liked to cosplay as a tatterdemalion, and duped some Oxford philosophers. The underlying fundamentals never change much. In one of the earliest and most prescient commentaries on the FTX affair, the podcaster Dwarkesh Patel interviewed the financial journalist Bethany McLean, whose book “The Smartest Guys in the Room” remains the definitive account of the Enron debacle. This was in December, 2022, when details of Bankman-Fried’s purported caper were still scarce. For his first question, Patel asked McLean, “What are the odds that S.B.F. read ‘The Smartest Guys in the Room’ and just followed it as a playbook?” McLean laughed and said she loved the idea, then thought for a moment and took it back. “I actually think that, even if he had read the book, it would never have occurred to him that there was a similarity, because self-delusion is such a strong component of all these stories,” she said. “It’s very rare that you have one of the characters at the heart of this who actually understands what they’re doing and understands that they’re moving over into the dark side and thinks about the potential repercussions of this and chooses this path anyway. That’s usually not the way these stories go.”

The trial’s courtroom is on the twenty-sixth and top floor of a high-rise, with reinforced glass, purplish tapestries, and tall sash-curtained windows that frame sweeping views of downtown—as if the government wanted to assert a level of grandeur and authority at least roughly in line with those enjoyed by its finest potential defendants. Three wooden pews in the rear of the oddly elongated chamber are reserved for the gallery. Before the strictly shushed proceedings began, a reporter from The Verge asked after Taco’s government name, a faux pas that threw into sharp relief the differences between the world of generic trust and its crypto parallel. First, he said the government considered him dead. Then he said, “Stephen Smith.”

“Really?” the reporter asked.

“No,” he said.

The prosecution’s strategy relied on two things: that Bankman-Fried had been unable to keep his mouth shut, and that three of Bankman-Fried’s closest friends, who were also his steadfast lieutenants, his roommates, and, according to the prosecution, his co-conspirators, had turned government’s witness. The marquee attraction was Caroline Ellison, Bankman-Fried’s subordinate and his thwarted accomplice in an immiserating situationship. She had inherited the role of co-C.E.O. at Alameda when Bankman-Fried applied himself to FTX full time, and the two had had an on-and-off affair he kept largely hidden. The previous day, Ellison had been asked to identify the defendant. The Times reported that it took her ten seconds, but in the sealed terrarium of the courtroom time took on an elastic quality, and nobody could agree on exactly how long it had taken. (Estimates ranged from five seconds to a full minute.) In the outside world, people wondered why Ellison had found the task so difficult. Was it her anxiety? Was it because she’d never before seen Bankman-Fried in formal clothes, or because she didn’t recognize him with a new jail haircut, with his sideburns shaved up to the temples? The courtroom sketches involved a dramatic foreshortening of the premises, such that it looked as though the judge, witness, and defendant were practically on top of one another. In the venue itself, though, she was separated from her onetime lover and alleged puppet master by perhaps thirty feet of lawyers and screens. From the gallery, our view of him was similarly occluded. We could more easily see Bankman-Fried’s parents, two small, square-jawed figures with close-cropped hair, minor bruxism, and large legal pads.

Ellison entered from the rear at a brisk clip, with thick glasses and a business suit, her hair damp and parted with enough severity to give her a thin white stripe of exposed scalp. She explained her dawning realization, over the spring of 2022, that Alameda Research probably didn’t have sufficient funds to pay down what it had borrowed. After the crypto market entered a downturn that spring and many of Alameda’s loans were recalled, she testified, Bankman-Fried directed her to repay the lenders with money that could only have been drawn from their sixty-five-billion-dollar line of credit with FTX—that is, with assets that customers believed safe in the exchange’s custody. The situation hadn’t thus far seemed imminently catastrophic: at one point, Bankman-Fried asked her to assess the likelihood of a destabilizing bank run, an outcome to which she assigned a probability of only thirteen per cent—not great, certainly, but not yet a disastrously foregone conclusion. When Bankman-Fried put another three billion dollars into further venture investments, however, that probability rose to about one in five, which Ellison considered “extremely concerning.” By the summer, as Alameda’s outstanding loans from various crypto desks were recalled and its debt to unassuming customers grew, she became “very stressed out.” Still, she said, she thought there was some chance Bankman-Fried might “be able to fix things somehow”—that he could simply raise more money to fill the abyss that was starting to yawn beneath their feet.

Her testimony culminated in the story of how she had, at Bankman-Fried’s urging to come up with “alternative ways of presenting the information,” prepared seven different versions of their balance sheet for inquiring counterparties. As Bloomberg’s Matt Levine put it in the next day’s column, “If you prepare a balance sheet for a lender and your boss says ‘why don’t we present this information in a different way,’ you probably need a lawyer. If you prepare SEVEN BALANCE SHEETS and your boss is like ‘let’s go with Alternative 7’ then one of you is going to prison absolutely forever.” Bankman-Fried, on the prosecutor’s account, was brazenly deceptive in his public assurances during this time: a few weeks after the balance-sheet triage, he tweeted, “Backstopping customer assets should always be primary. Everything else is secondary.”

Ellison’s testimony might have had the most lurid appeal, but her basic point—that Bankman-Fried had been kept apprised of Alameda’s financial problems, had ordered the use of customer funds to compensate, and had deliberately misled customers, investors, lenders, the general public, and Congress about it—was reinforced by similar testimony from Gary Wang, FTX’s C.T.O., and Nishad Singh, the company’s director of engineering. Wang and Singh had, at various points, written the code that gave Alameda’s accounts on FTX carte blanche to carry negative balances, and, as 2022 wore on, all of them had come to understand that the feature had been abused. By September, company meetings had devolved into open confrontation, when Singh told Bankman-Fried, poolside on their penthouse terrace, that the promiscuous distribution of customer funds had to stop.

The rest of the fall played out exactly as Ellison had feared. By early October, Ellison said, Alameda’s debt to FTX customers totalled fourteen or fifteen billion dollars. (The prosecutor, showing the numbers, promised the judge that “this is the last spreadsheet”; he dryly noted that “Microsoft stock must be plunging.”) By then, Ellison said, there was no way to repay their debts unless crypto went way up or they sold FTX equity. Bankman-Fried thought that Mohammed bin Salman, the Saudi crown prince, might be courted for a capital infusion. But the game was over. Bankman-Fried and Ellison were barely on speaking terms at that point, but he told her the whole problem was due to her failure to hedge their positions properly, or at all. This ignored his own tendency to compulsive spending—a habit that included, among many other things, an investment of hundreds of millions of dollars, during the time the company was in trouble, in Modulo Capital, an Alameda competitor run by another former flame.

In a Signal group chat dated November 6th, Singh noted that customers had withdrawn their funds at the rate of a hundred and twenty million dollars in the past hour. Ellison replied with a sad-face emoticon, which seemed to have understated her emotional volatility. She told the court that this had been “over all the worst week of my life.” At the end of the day, when Ellison described the overwhelming sense of relief she’d felt when she no longer had to anticipate the future with dread, she broke down in sniffly tears. In person, she seemed distraught enough. In the almost comically unflattering courtroom sketch, which caught her clutching a crumpled tissue to her face, she looked like a marmot with a head cold.

It wasn’t easy to argue with the prosecution’s evidence, which most onlookers have described as “overwhelming.” It remained unclear, however, what the jurors—at least one of whom seemed to be sleeping at any given time—made of it all. For one thing, the numbers involved—eight billion here, three billion there, 4.1 billion elsewhere, two hundred and fifty million requested casually over Signal chat to be sent to an indecipherable alphanumeric string—gave the impression that they were talking about Monopoly money. The sums never felt quite real; at one point, a line item in an exhibit showed Alameda liabilities to FTX as -19,003,461,370. And, in some sense, a lot of it was Monopoly money. Michael Lewis’s attitude, which has not endeared him to crypto enthusiasts, is, at least implicitly, that we need not feel overly bad for gambling addicts sitting down to the get-rich-quick tables at a dodgy overseas casino. The prosecution never introduced the sort of witness-victim one might expect—a little old lady in Peoria who lost her retirement savings to a mustache-twirling villain. Instead, they produced a well-heeled French cocoa trader. The jurors might well wonder why such figures deserve to be the object of our pity.

The other major problem with Ellison as a witness is that she needed to come off as independent enough, and thus credible enough, to know what was going on and to know it was wrong—because if she didn’t know it was wrong it might equally well be the case that Bankman-Fried didn’t, either—but not so independent that the jury might believe the defense’s version of the story: that she was the source of the problem, because she squandered huge sums on bad trades. This coalesced with another issue: If she was so disturbed by what was going on, why didn’t she quit much earlier? After all, her co-C.E.O. at Alameda, Sam Trabucco, had bolted for good in August. Her asymmetrical relationship with Bankman-Fried presumably played some role; as the host of “The CPR Show,” which covers N.F.T.s and crypto, told me the next morning in line, “If I’d been in love with my boss when I was twenty-five and he’d asked me to murder somebody, I probably would have.”

The lead prosecutor seemed to draw generally positive reviews; her parents were regularly in the courtroom, sitting across from the Bankman-Frieds, and seemed proud of their precocious daughter. The same could not be said for Bankman-Fried’s lawyer, whose avant-garde sense of chronology seemed like something out of “Cloud Atlas.” “We’re going to come back to June, we’re going to come back to September,” he said at one point. He struggled with some of the basics of the intertwined FTX and Alameda ledgers—which were, in his defense, confusing by design—and seemed hemmed in by admissibility constraints. In his cross-examination of Ellison, he seemed poised to reference an earlier moment of tumult at Alameda, when about four million dollars of cryptocurrency went missing. It later reappeared, which seemed to vindicate Bankman-Fried’s ability to play fast and loose with funds and have it all turn out O.K. When he brought this up, the prosecutors objected, and the inquiry was petulantly dropped.

One reporter, working on an elaborate courtroom sketch of his own, said, to no one in particular, of the tenuousness of this argument, “That is a slender reed, my dudes.” And, for the present purposes, it was. But it gestured to a broader context. The judge had also ruled that the defense was not permitted to bring up the matter of a five-hundred-million-dollar venture investment that Bankman-Fried had made in an A.I. startup called Anthropic. As with many of Bankman-Fried’s transactions, this was a little incestuous: Anthropic was co-founded by Dario Amodei and his sister Daniela, the wife of Holden Karnofsky, an influential figure in the world of effective altruism—the loose cohort of rigorous, utilitarian-inflected philanthropic thinkers with which Bankman-Fried was associated. But the bet seemed to validate Bankman-Fried’s calculations of expected value. Anthropic could soon be valued at as much as thirty billion dollars. If only the world had been a little patient, the missing customer money might have been handily recouped. Here was a clear Enron parallel: Enron Broadband, a proto-streaming initiative, was essentially a premature version of Netflix; in a slightly different world, the company might have pulled it off and no one would have been the wiser.

These kinds of branching paths in no way affected the trial. Stealing money and lying about it is stealing money and lying about it, irrespective of one’s ability to pay it back. But, if one takes seriously Lewis’s portrait of Bankman-Fried’s self-image as a kind of deadpan bettor in the probabilistic casino of the universe, one can’t help but feel like the existence of the trial, as necessary as it is, seems a little arbitrary. There are likely plenty of people who do similar things and never get caught, or stay just this side of murky legal lines. Perhaps it is worth entertaining the counterfactual in which the bank run was forestalled, the Anthropic investment went to the moon, customers were whole. Nothing that Bankman-Fried seems to have done was right, or fair, or justifiable, but one can nevertheless imagine, as Lewis asks us to, that, by Bankman-Fried’s own megalomaniacally “rational” lights, it might have seemed that way. If he came by anything honestly, it was his capacity for this kind of rationalization. Bankman-Fried’s mother wrote in “Beyond Blame,” a contribution to a 2013 issue of Boston Review, that our “blame fest” has given us nothing “except the guilty pleasure of reproaching others for acts that, but for the grace of God, or luck, or social or biological forces, we might well have committed ourselves.”

Over the course of the trial, the jurors were largely spared an extended discussion of the doctrines of effective altruism, although Ellison did mention at one point that Bankman-Fried’s pure utilitarianism left him unbothered by the ethical strictures—don’t lie, don’t steal—with which ordinary people are saddled. For what it’s worth, most effective altruists express a belief in utilitarianism “with side constraints”—that is, their aspirations to do the greatest good for the greatest number tend to be leavened with the normie morality of at least the major Commandments, if not all ten. Effective altruism was, however, present in style if not in substance. Even the most winning E.A.s—a group that includes Nishad Singh, someone I personally liked very much when I interviewed him eighteen months ago—are the most pedantic people alive. They believe they speak the mathematically precise language of the knowable universe. When Singh took the stand and banged on about things like auto-deleveraging events and liquidity backstop providers, the judge, his patience with E.A. specificity wearing thin, cut him off: “You stick to answering the questions that are asked, O.K.?”

Bankman-Fried’s defense, in Lewis’s account, is that it couldn’t have been fraud because he wasn’t paying enough attention. Lewis has been criticized for seeming to advocate on Bankman-Fried’s behalf, and he has certainly said some odd things—that, to his mind, Bankman-Fried doesn’t engage in “outright lying,” for example—that have not held up particularly well. In my view, the book is best read as a portrait of self-delusion. As Bethany McLean put it in her interview about Enron, “It’s not bad people setting out to do bad things—but it’s human beings at first convincing themselves even that they’re doing the right thing, and then ending up in a situation that they never meant to be in.” The defense has attempted to show, in its clumsy way, that fantasy prevailed for everybody all the way down. It was only at the very last second that Ellison seemed to rule out the possibility of a miraculous rescue. Singh testified that he didn’t know the extent of the financial problems until September, 2022, which seems far-fetched, but E.A.s are very good at not knowing things they don’t care to know about. But, if Singh didn’t know until very late in the game, the idea, by proxy, is that Bankman-Fried might have been similarly deluded, not in spite of his intelligence but in part because of it. Singh might have been “concerned about frivolous spending,” but in October, 2022, he went out and bought himself a $3.7-million house on Orcas Island. Although he said on the stand that he had forfeited the property, calling the purchase “egregious,” it seems as though, even after he knew that every dollar spent had effectively come from a customer, he did not—and perhaps could not—update his own self-image to that of a thief.

On the first day of Sam Bankman-Fried’s testimony, the mood was feverish. The list for the door was closed by 3:15 A.M. Michael Lewis ended up in one of the overflow rooms (rumor had it that there were eight), gamely signing books. (After lunch, he talked his way into the courtroom; one of the court security officers was a fan of his podcast.) Bankman-Fried appeared, on the low-resolution screens of the overflow room, pancaked in white makeup; with his boxy helmet of an amateur haircut, he looked like a Minecraft character. His initial testimony was held in the absence of the jury, to discuss the admissibility of certain evidence. It was probably in the interest of the defense that the jury didn’t see it. The prosecutor was at her most confident and incisive, and the audience thrilled to her cold, mechanical bloodlust. Ellison had testified that her work with Bankman-Fried had turned her into a worse person, and he seemed to have an effect on the overflow room I was in as well, which now seemed disorderly in a faintly menacing way. There were minor scuffles over the good seats.

The narrative that Bankman-Fried tried to tell was familiar to readers of the Lewis book. In his attempts to set up a cryptocurrency exchange that did not socialize losses—where, in other words, customers would never be responsible for the bad trades of other customers—he allowed Alameda to lose money instead. Then, in part because Ellison was inadequately hedged, Alameda lost a lot of money. But Bankman-Fried, in part because of coding issues and in part because of the lack of good accounting and adequate risk management, never really knew how much had been lost until it was too late to do anything about it. The company had been too successful and had grown too fast. The crux of the issue, in the end, was not what had happened—everybody agreed that customer funds had been spent—but what Bankman-Fried knew when. Bankman-Fried claimed a kind of perpetual, well-meaning cluelessness. The government found this ridiculous. It showed a series of devastating time lines that laid out, in precise detail, six different points in 2021 and 2022 when Bankman-Fried was faced with the explicit choice “to come clean or to double down.” In each case, he doubled down. This was, the government argued, a guy who lied constantly. He lied to investors, he lied to customers, he lied to George Stephanopoulos on “Good Morning America,” he lied to Congress. His decision to testify allowed the government to tell the jurors that Bankman-Fried had also lied to them.

The best defense of Elizabeth Holmes was that she was merely an easy scapegoat for the move-fast-break-things mentality of Silicon Valley. Some of the things that Bankman-Fried did were not wholly uncommon in the cryptocurrency industry. The prosecutor made some gestures in the direction of the unfairness of his scheme—that, on FTX, Alameda enjoyed inequitable advantages over the poor-schmuck retail investor who lost his “nest egg.” This might have been true, but it felt halfhearted. Nevertheless, the trial found its ritual significance. At one point, early in the closing arguments, the prosecutor referred to how “smart” Bankman-Fried was. The same thread had run through the disappointed reactions to the Lewis book—that Lewis seemed so impressed by Bankman-Fried’s alleged raw intelligence that the bright young man simply couldn’t have been a run-of-the-mill fraud. He was instead just too recklessly brilliant for his own good. This seems true insofar as he, like everybody else in his situation, assumed he was clever enough to escape from the trap of his own design. He was never delusional about his companies’ financial situation; he was delusional about his ability to bail himself out. Bankman-Fried’s lawyer claimed that the government had made a “monster” out of him. They had used his hair and his clothes and his celebrity pals to make him into someone the jury would dislike, because “every movie needs a villain. And, let’s face it, an awkward high-school math nerd doesn’t look villainous.” But the government, it seemed to me, had just portrayed Bankman-Fried as a petty, greedy tyrant. He wasn’t a monster. He just thought he was smarter than everybody else.

On Thursday evening, after less than five hours of deliberation, the jury found Bankman-Fried guilty on all seven counts. He faces up to a hundred and ten years in prison. He will be sentenced at the end of March. It remains unclear, though, what the verdict will achieve aside from retribution. Deterrence is the obvious idea. As McLean has pointed out, both in her interview and in the afterword to the tenth-anniversary edition of her book, Jeffrey Skilling’s sentence, alongside new financial regulation, held out such promise. But very few people look on at grifters and recognize themselves. As she put it, “We wrote a piece for Fortune in which we said that the entire world has changed, now that corporate executives are put on high alert that behavior in the gray area will no longer be tolerated and that it will be aggressively prosecuted. And this was spring of 2006 and the events that caused the global financial crisis were pretty well under way. It didn’t do much to prevent the global financial crisis. Enron’s jail time didn’t do anything to prevent it. Elizabeth Holmes doesn’t seem to have done anything to change what S.B.F. was doing. So I just, I’m not sure—I’m sure a psychologist or somebody who specializes in studying white-collar crime could probably make an argument that refutes everything I said and that shows that it has had a deterrent effect. But I just don’t think that people who get themselves into this situation consciously think, This is what I’m doing.” ♦

An earlier version of this article misstated the valuation of Anthropic.