Sam Bankman-Fried’s $30 million Bahamas penthouse looks like a dorm after the students have left for winter break. The dishwasher is full. Towels are piled in the laundry room. Bat streamers from a Halloween party are still hanging from a doorway. Two boxes of Legos sit on the floor of one bedroom. And then there are the shoes—dozens of sneakers and heels piled in the foyer, left behind by employees who fled the island of New Providence last month when his cryptocurrency exchange FTX imploded.
“It’s been an interesting few weeks,” Bankman-Fried says in a chipper tone as he greets me. It’s a muggy Saturday afternoon, eight days after FTX filed for bankruptcy. He’s shoeless, in white gym socks, a red T-shirt and wrinkled khaki shorts. His standard uniform.
This isn’t part of the typical tour Bankman-Fried gave to the many reporters who came to tell the tale of the boy-genius-crypto-billionaire who slept on a beanbag chair next to his desk and only got rich so he could give it all away, and it’s easy to see why. The apartment is at the top of one of the luxury condo buildings that border a marina in a gated community called Albany. Outside, deckhands buff the stanchions of a 200-foot yacht owned by a fracking billionaire. A bronze replica of Wall Street’s Charging Bull statue stands on the lawn, which is as manicured as the residents. I feel like I’ve crash-landed on an alien planet populated solely by the very rich and the people who work for them.
Bankman-Fried leads me down a marble-floored hallway to a small bedroom, where he perches on a plush brown couch. Always known for being jittery, he taps his foot so hard it rattles a coffee table, smacks gum and rubs his index finger with his thumb like he’s twirling an invisible fidget spinner. But he seems almost cheerful as he explains why he’s invited me into his 12,000-square-foot bolthole, against the advice of his lawyers, even as investigators from the US Department of Justice probe whether he used customers’ funds to prop up his hedge fund, a crime that could send him to prison for years. (Spoiler alert: It sure looks like he did.)
“What I’m focusing on is what I can do, right now, to try and make things as right as possible,” Bankman-Fried says. “I can’t do that if I’m just focused on covering my ass.”
But he seems to be doing just that, with me here and all along the apology tour he’ll later embark on, which will include a video appearance at a New York Times conference and an interview on Good Morning America. He’s been trying to blame his firm’s failure on a hazy combination of comically poor bookkeeping, wildly misjudged risks and complete ignorance of what his hedge fund was doing. In other words, an alumnus of both MIT and the elite Wall Street trading firm Jane Street is arguing that he was just dumb with the numbers—not pulling a conscious fraud. Talking in detail to journalists about what’s certain to be the subject of extensive litigation seems like an unusual strategy, but it makes sense: The press helped him create his only-honest-man-in-crypto image, so why not use them to talk his way out of trouble?
He doesn’t say so, but one reason he might be willing to speak with me is that I’m one of the reporters who helped build him up. After spending two days at FTX’s offices in February, I flew past the bright red flags at his company—its lack of corporate governance, the ties to his Alameda Research hedge fund, its profligate spending on marketing, the fact that it operated largely outside US jurisdiction. I wrote a story focused on whether Bankman-Fried would follow through on his plans to donate huge sums to charity and his connections to an unusual philanthropic movement called effective altruism.
It wasn’t the most embarrassingly puffy of the many puff pieces that came out about him. (“After my interview with SBF, I was convinced: I was talking to a future trillionaire,” one writer said in an article commissioned by a venture capital firm.) But my tone wasn’t entirely dissimilar. “Bankman-Fried is a thought experiment from a college philosophy seminar come to life,” I wrote. “Should someone who wants to save the world first amass as much money and power as possible, or will the pursuit corrupt him along the way?” Now it seems pretty clear that a better question would’ve been whether the business was a scam from the start.
I tell Bankman-Fried I want to talk about the decisions that led to FTX’s collapse, and why he took them. Earlier in the week, in late-night DM exchanges with a Vox reporter and on a phone call with a YouTuber, he made comments that many interpreted as an admission that everything he said was a lie. (“So the ethics stuff, mostly a front?” the Vox reporter asked. “Yeah,” Bankman-Fried replied.) He’d spoken so cynically about his motivations that to many it seemed like a comic book character was pulling off his mask to reveal the villain who’d been hiding there all along.
I set out on this visit with a different working theory. Maybe I was feeling the tug of my past reporting, but I still didn’t think the talk about charity was all made up. Since he was a teenager, Bankman-Fried has described himself as utilitarian—following the philosophy that the correct action is the one likely to result in the greatest good for the greatest number of people. He said his endgame was making and donating enough money to prevent pandemics and stop runaway artificial intelligence from destroying humanity. Faced with a crisis, and believing he was the hero of his own sci-fi movie, he might’ve thought it was right to make a crazy, even illegal, gamble to save his company.
To be clear, if that’s what happened, it’s the logic of a megalomaniac, not a martyr. The money wasn’t his to gamble with, and “the ends justify the means” is a cliché of bad ethics. But if it’s what he believed, he might still think he’d made the right decision, even if it didn’t work out. It seemed to me that’s what he meant when he messaged Vox, “The worst quadrant is sketchy + lose. The best is win + ???” I want to probe that, in part because it might get him to talk more candidly about what had happened to his customers’ money.
I decide to approach the topic gingerly, on terms I think he’ll relate to, as it seems he’s in less of a crime-confess-y mood. He’s said he likes to evaluate decisions in terms of expected value—the odds of success times the likely payoff—so I begin by asking: “Should I judge you by your impact, or by the expected value of your decision?”
“When all is said and done, what matters is your actual realized impact. Like, that’s what actually matters to the world,” he says. “But, obviously, there’s luck.”
That’s the in I’m looking for. For the next 11 hours—with breaks for fundraising calls and a very awkward dinner—I try to get him to tell me exactly what he meant. He denies that he’s committed fraud or lied to anyone and blames FTX’s failure on his sloppiness and inattention. But at points it seems like he’s saying he got unlucky, or miscalculated the odds.
Bankman-Fried tells me he’s still got a chance to raise $8 billion to save his company. He seems delusional, or committed to pretending this is still an error he can fix, and either way, the few supporters remaining at his penthouse seem unlikely to set him straight. The grim scene reminds me a bit of the end of Scarface, with Tony Montana holed up in his mansion, semi-incoherent, his unknown enemies sneaking closer. But instead of mountains of cocaine, Bankman-Fried is clinging to spreadsheet tabs filled with wildly optimistic cryptocurrency valuations.
Think of FTX like an offshore casino. Customers sent in money, then gambled on the price of hundreds of cryptocurrencies—not just Bitcoin or Ether, but more obscure coins. In crypto slang, the latter are called shitcoins, because almost no one knows what they’re for. But in the past few years, otherwise respectable people, from retired dentists to heads of state, convinced themselves that these coins were the future of finance. Or at least that enough other people might think so to make the price go up. Bankman-Fried’s casino was growing so fast that earlier this year some of Silicon Valley’s top venture capitalists invested in it at a $32 billion valuation.
The problem surfaced last month. After a rival crypto-casino kingpin raised concerns about FTX on Twitter, customers rushed to cash in their chips. But when Bankman-Fried’s casino opened the vault, their money wasn’t there. According to multiple news reports citing people familiar with the matter, it had been secretly lent to Bankman-Fried’s hedge fund, which had lost it in some mix of bad bets, insane spending and perhaps something even sketchier. John Ray III, the lawyer who’s now chief executive officer of the bankrupt exchange, has alleged in court that FTX covered up the loans using secret software.
Bankman-Fried denies this again to me. Returning to the framework of expected value, I ask him if the decisions he made were correct.
“I think that I’ve made a lot of plus-EV decisions and a few very large boneheaded decisions,” he says. “Certainly in retrospect, those very large decisions were very bad, and may end up overwhelming everything else.”
The chain of events, in his telling, started about four years ago. Bankman-Fried was in Hong Kong, where he’d moved from Berkeley, California, with a small group of friends from the effective-altruism community. Together they ran a successful startup crypto hedge fund, Alameda Research. (The name itself was an early example of his casual attitude toward rules—it was chosen to avoid scrutiny from banks, which frequently closed its accounts. “If we named our company like, Shitcoin Daytraders Inc., they’d probably just reject us,” Bankman-Fried told a podcaster in 2021. “But, I mean, no one doesn’t like research.”)
The fund had made millions of dollars exploiting inefficiencies across cryptocurrency exchanges. (Ex-employees, even those otherwise critical of Bankman-Fried, have said this is true, though some have said Alameda then lost some of that money because of bad trades and mismanagement.) Bankman-Fried and his friends began considering starting their own exchange—what would become FTX.
The way Bankman-Fried later described this decision reveals his attitude toward risk. He estimated there was an 80% chance the exchange would fail to attract enough customers. But he’s said one should always take a bet, even a long-shot one, if the expected value is positive, calling this stance “risk neutral.” But it actually meant he would take risks that to a normal person sound insane. “As an individual, to make a bet where it’s like, ‘I’m going to gamble my $10 billion and either get $20 billion or $0, with equal probability,’ would be madness,” Rob Wiblin, host of an effective-altruism podcast, said to Bankman-Fried in April. “But from an altruistic point of view, it’s not so crazy.”
“Completely agree,” Bankman-Fried replied. He told another interviewer that he’d make a bet described as a chance of “51% you double the earth out somewhere else, 49% it all disappears.”
Bankman-Fried and his friends jump-started FTX by having Alameda provide liquidity. It was a huge conflict of interest. Imagine if the top executives at an online poker site also entered its high-stakes tournaments—the temptation to cheat by peeking at other players’ cards would be huge. But Bankman-Fried assured customers that Alameda would play by the same rules as everyone else, and enough people came to trade that FTX took off. “Having Alameda provide liquidity on FTX early on was the right decision, because I think that helped make FTX a great product for users, even though it obviously ended up backfiring,” Bankman-Fried tells me.
Part of FTX’s appeal was that it was mostly a derivatives exchange, which allowed customers to trade “on margin,” meaning with borrowed money. That’s a key to his defense. Bankman-Fried argues no one should be surprised that big traders on FTX, including Alameda, were borrowing from the exchange, and that his fund’s position just somehow got out of hand. “Everyone was borrowing and lending,” he says. “That’s been its calling card.” But FTX’s normal margin system, crypto traders tell me, would never have permitted anyone to accumulate a debt that looked like Alameda’s. When I ask if Alameda had to follow the same margin rules as other traders, he admits the fund did not. “There was more leeway,” he says.
That wouldn’t have been so important had Alameda stuck to its original trading strategy of relatively low-risk arbitrage trades. But in 2020 and 2021, as Bankman-Fried became the face of FTX, a major political donor and a favorite of Silicon Valley, Alameda faced more competition in that market-making business. It shifted its strategy to, essentially, gambling on shitcoins.
As Caroline Ellison, then Alameda’s co-CEO, explained in a March 2021 post on Twitter: “The way to really make money is figure out when the market is going to go up and get balls long before that,” she wrote, adding that she’d learned the strategy from the classic market-manipulation memoir, Reminiscences of a Stock Operator. Her co-CEO said in another tweet that a profitable strategy was buying Dogecoin because Elon Musk tweeted about it.
The reason they were bragging about what sounded like a high schooler’s tactics was that it was working better than anyone knew. When we spoke in February 2022, Bankman-Fried told me that Alameda had made $1 billion the previous year. He now says that was Alameda’s arbitrage profits. On top of that, its shitcoins gained tens of billions of dollars of value, at least on paper. “If you mark everything to market, I do believe at one point my net worth got to $100 billion,” Bankman-Fried says.
Any trader would know this wasn’t nearly as good as it sounded. The large pile of tokens couldn’t be turned into cash without crashing the market. Much of it was even made of tokens that Bankman-Fried and his friends had spun up themselves, such as FTT, Serum or Maps—the official currency of a nonsensical crypto-meets-mapping app—or were closely affiliated with, like Solana. While Bankman-Fried acknowledges the pile was worth something less than $100 billion—maybe he’d mark it down a third, he says—he maintains that he could have extracted quite a lot of real money from his holdings.
But he didn’t. Instead, Alameda borrowed billions of dollars from other crypto lenders—not FTX—and sunk them into more crypto bets. Publicly, Bankman-Fried presented himself as an ethical operator and called for regulation to rein in crypto’s worst excesses. But through his hedge fund, he’d actually become the market’s most degenerate gambler. I ask him why, if he really thought he could sell the tokens, he didn’t. “Why not, like, take some risk off?”
“OK. In retrospect, absolutely. That would’ve been the right, like, unambiguously the right thing to do,” he says. “But also it was just, like, hilariously well-capitalized.”
Near the peak of the great shitcoin boom, in April 2022, FTX hosted a lavish conference at a resort and casino in Nassau. It was Bankman-Fried’s coming out party. He got to share the stage with quarterback Tom Brady. Also there: former Prime Minister Tony Blair and ex-President Bill Clinton, who extended a fatherly hand when the young crypto executive seemed nervous. The author Michael Lewis, who’s working on a book about Bankman-Fried, praised him in a fawning interview onstage. “You’re breaking land speed records. And I don’t think people are really noticing what’s happened, just how dramatic the revolution has become,” Lewis said, asking when crypto would take over Wall Street.
The next month, the crypto crash began. It started when a popular set of coins called Terra and Luna collapsed, wiping out $60 billion. Terra and Luna were almost openly a Ponzi scheme, but some of the biggest crypto funds had invested in them with borrowed money and went bankrupt. This made the lenders who’d lent billions of dollars to Alameda nervous. They asked Alameda to repay the loans, with real money. It needed billions of dollars, fast, or it would go bust.
There are two different versions of what happened next. Two people with knowledge of the matter told me that Ellison, by then the sole head of Alameda, had told her side of the story to her staff amid the crisis. Ellison said that she, Bankman-Fried and his two top lieutenants—Gary Wang and Nishad Singh—had discussed the shortfall. Instead of admitting Alameda’s failure, they decided to use FTX customer funds to cover it, according to the people. If that’s true, all four executives would’ve knowingly committed fraud. (Ellison, Wang and Singh didn’t respond to messages seeking comment.)
When I put this to Bankman-Fried, he screws up his eyes, furrows his eyebrows, puts his hands in his hair and thinks for a few seconds.
“So, it’s not how I remember what happened,” Bankman-Fried says. But he surprises me by acknowledging that there had been a meeting, post-Luna crash, where they debated what to do about Alameda’s debts. The way he tells it, he was packing for a trip to DC and “only kibitzing on parts of the discussion.” It didn’t seem like a crisis, he says. It was a matter of extending a bit more credit to a fund that already traded on margin and still had a pile of collateral worth way more than enough to cover the loan. (Although the pile of collateral was largely shitcoins.)
“That was the point at which Alameda’s margin position on FTX got, well, it got more leveraged substantially,” he says. “Obviously, in retrospect, we should’ve just said no. I sort of didn’t realize then how large the position had gotten.”
“You were all aware there was a chance this would not work,” I say.
“That’s right,” he says. “But I thought that the risk was substantially smaller.”
I try to imagine what he could’ve been thinking. If FTX had liquidated Alameda’s position, the fund would’ve gone bankrupt, and even if the exchange didn’t take direct losses, customers would’ve lost confidence in it. Bankman-Fried points out that the companies that lent money to Alameda might have failed, too, causing a hard-to-predict cascade of events.
“Now let’s say you don’t margin call Alameda,” I posit. “Maybe you think there’s like a 70% chance everything will be OK, it’ll all work out?”
“Yes, but also in the cases where it didn’t work out, I thought the downside was not nearly as high as it was,” he says. “I thought that there was the risk of a much smaller hole. I thought it was going to be manageable.”
Bankman-Fried pulls out his laptop (an Acer Predator) and opens a spreadsheet to show what he meant. It’s similar to the balance sheet he reportedly showed investors when he was seeking a last-minute bailout, which he says consolidated FTX and Alameda’s positions because by then the fund had defaulted on its debt. On one line—labeled “What I *thought*”—he lists $8.9 billion in debts and way more than enough money to pay them: $9 billion in liquid assets, $15.4 billion in “less liquid” assets and $3.2 billion in “illiquid” ones. He tells me this was more or less the position he was considering when he had the meeting with the other executives.
“It looks naively to me like, you know, there’s still some significant liabilities out there, but, like, we should be able to cover it,” he says.
“So what’s the problem, then?”
Bankman-Fried points to another place on the spreadsheet, which he says shows the actual truth of the situation at the time of the meeting. This one shows similar numbers, but with $8 billion less liquid assets.
“What’s the difference between these two rows here?” he asks.
“You didn’t have $8 billion in cash that you thought you had,” I say.
“That’s correct. Yes.”
“You misplaced $8 billion?” I ask.
“Misaccounted,” Bankman-Fried says, sounding almost proud of his explanation. Sometimes, he says, customers would wire money to Alameda Research instead of sending it directly to FTX. (Some banks were more willing to work with the hedge fund than the exchange, for some reason.) He claims that somehow, FTX’s internal accounting system double-counted this money, essentially crediting it to both the exchange and the fund.
That still doesn’t explain why the money was gone. “Where did the $8 billion go?” I ask.
To answer, Bankman-Fried creates a new tab on the spreadsheet and starts typing. He lists Alameda and FTX’s biggest cash flows. One of the biggest expenses is paying a net $2.5 billion to Binance, a rival, to buy out its investment in FTX. He also lists $250 million for real estate, $1.5 billion for expenses, $4 billion for venture capital investments, $1.5 billion for acquisitions and $1 billion labeled “fuckups.” Even accounting for both firms’ profits, and all the venture capital money raised by FTX, it tallies to negative $6.5 billion.
Bankman-Fried is telling me that the billions of dollars customers wired to Alameda is gone simply because the companies spent way more than they made. He claims he paid so little attention to his expenses that he didn’t realize he was spending more than he was taking in. “I was real lazy about this mental math,” the former physics major says. He creates another column in his spreadsheet and types in much lower numbers to show what he thought he was spending at the time.
It seems to me like he is, without saying it exactly, blaming his underlings for FTX’s failure, especially Ellison, the head of Alameda. The two had dated and lived together at times. She was part of Bankman-Fried’s Future Fund, which was supposed to distribute FTX and Alameda’s earnings to effective-altruist-approved causes. It seems unlikely she would’ve blown billions of dollars without asking. “People might take, like, the TLDR as, like, it was my ex-girlfriend’s fault,” I tell him. “That is sort of what you’re saying.”
“I think the biggest failure was that it wasn’t entirely clear whose fault it was,” he says.
Bankman-Fried tells me he has to make a call. After a while, the sun goes down and I’m hungry. I’m allowed to join a group of Bankman-Fried’s supporters for dinner, as long as I don’t mention their names.
With the curtains drawn, the living room looks considerably less grand than it does in pictures. I’ve been told that FTX employees gathered here amid the crisis, while Bankman-Fried worked in another apartment. Addled by stress and sleep deprivation, they wept and hugged one another. Most didn’t say goodbye as they left the island, one by one. Many flew back to their childhood homes to be with their parents.
The supporters at the dinner tell me they feel like the press has been unfair. They say that Bankman-Fried and his friends weren’t the polyamorous partiers the tabloids have portrayed and that they did little besides work. Earlier in the week, a Bahamian man who’d served as FTX’s round-the-clock chauffeur and gofer also told me the reports weren’t true. “People make it seem like this big Wolf of Wall Street thing,” he said. “Bro, it was a bunch of nerds.”
By the time I finish my plate of off-the-record rice and beans, Bankman-Fried is free again. We return to the study. He’s barefoot now, having balled up his gym socks and stuffed them behind a couch cushion. He lies on the couch, his computer on his lap. The light from the screen casts shadows of his curls on his forehead.
I notice a skin-colored patch on his arm. He tells me it’s a transdermal antidepressant, selegiline. I ask if he’s using it as a performance enhancer or to treat depression. “Nothing’s binary,” he says. “But I’ve been borderline depressed for my whole life.” He adds that he also sometimes takes Adderall—“10 milligrams at a time, a few times a day”—as did some of his colleagues, but that talk of drug use is overblown. “I don’t think that was the problem,” he says.
I tell Bankman-Fried my theory about his motivation, sidestepping the question of whether he misappropriated customer funds. Bankman-Fried denies that his world-saving goals made him willing to take giant gambles. As we talk more, it seems like he’s saying he made some kind of bet but hadn’t calculated the expected value properly.
“I was comfortable taking the risk that, like, I may end up kind of falling flat,” he says, staring at his computer screen, where he had pulled up a game and was leading an army of cartoon knights and fairies into battle. “But what actually happened was disastrously bad and, like, no significant chance of that happening would’ve made sense to risk, and that was a fuckup. Like, that was a mass miscalculation in downside.”
I read Bankman-Fried a post by Will MacAskill, one of the founders of the effective-altruism movement. He recruited Bankman-Fried into it when he was a junior at MIT and this year had joined the board of Bankman-Fried’s Future Fund. On Nov. 11, MacAskill wrote on Twitter that Bankman-Fried had betrayed him. “For years, the EA community has emphasized the importance of integrity, honesty and the respect of common-sense moral constraints,” MacAskill wrote. “If customer funds were misused, then Sam did not listen; he must have thought he was above such considerations.”
Bankman-Fried closes his eyes and pushes his toes against one arm of the couch, clenching the other arm with his hands. “That’s not how I view what happened,” he says. “But I did fuck up. I think really what I want to say is, like, I’m really fucking sorry. By far the worst thing about this is that it will tarnish the reputation of people who are dedicated to doing nothing but what they thought was best for the world.” Bankman-Fried trails off. On his computer screen, his army casts spells and swings swords unattended.
I ask what he’d say to people who are comparing him to the most famous Ponzi schemer of recent times. “Bernie Madoff also said he had good intentions and gave a lot to charity,” I say.
“FTX was a legitimate, profitable, thriving business. And I fucked up by, like, allowing a margin position to get too big on it. One that endangered the platform. It was a completely unnecessary and unforced error, which like maybe I got super unlucky on, but, like, that was my bad.”
“It fucking sucks,” he adds. “But it wasn’t inherent to what the business was. It was just a fuckup. A huge fuckup.”
To me, it doesn’t really seem like a fuckup. Even if I believe that he misplaced and accidentally spent $8 billion, he’s already told me that Alameda had been allowed to violate FTX’s margin rules. This wasn’t some little technical thing. He was so proud of FTX’s margining system that he’d been lobbying regulators for it to be used on US exchanges instead of traditional safeguards. In May, Bankman-Fried himself said on Twitter that exchanges should never extend credit to a fund and put other customers’ assets at risk. He wrote that the idea an exchange would even have that discretion was “scary.” I read him the tweets and ask: “Isn’t that, like, exactly what you did, right around that time?”
“Yeah, I guess that’s kind of fair,” he says. Then he seems to claim that this was evidence the rules he was lobbying for were a good idea. “I think this is one of the things that would have stopped.”
“You had a rule on your platform. You didn’t follow it,” I say.
By now it’s past midnight, and—operating without the benefit of any prescription stimulants—I’m worn out. I ask Bankman-Fried if I can see the apartment’s deck before I leave. Outside, crickets chirp as we stand by the pool. The marina is dark, lit only by the spotlights of yachts. As I say goodbye, Bankman-Fried bites into a burger bun and starts talking about potential bailouts with one of his supporters. —With Annie Massa and Gillian Tan
©2022 Bloomberg L.P.
