How Three Arrows Capital Blew Up And Set Off A Crypto Contagion

That kind of steely optimism isn’t hard to find in a crowd that turned the typo “HODL” into a full-blown mantra for never selling. But Zhu wasn’t just any laser-eyed crypto trader. With his schoolmate Kyle Davies, he ran Three Arrows Capital, one of the largest crypto hedge funds in the world. With a few billion dollars under management, it was far from massive by Wall Street standards. But in digital assets, it was a heavyweight.

More than that, Zhu and Davies were an integral part of the crypto market’s densely woven web. Their fund was a venture investor in some of the best-known crypto startups and, in some cases, manager of their corporate treasuries. It was both an aggressive borrower from large lenders and a shareholder in some of those lenders. It was a corporate parent for other fledgling funds. The duo were influencers with a combined 610,000 Twitter followers, as well as brokers of deals and introductions.

Zhu first made his name in late 2018 by rightly calling the end of the last “crypto winter,” which saw the price of Bitcoin fall some 80%. So when Bitcoin fell from its peak of more than $68,000 this year, buffeted by rising interest rates that sent investors fleeing from risky assets, he remained sanguine. With borrowed cash, Three Arrows placed big bets that the cryptocurrency would rebound. Instead, the market kept sinking, one domino falling after another until the fund became the biggest domino of all. It started missing margin calls from the companies funding its trades in mid-June and declared bankruptcy on July 1, as Bitcoin traded below $20,000.

In the July 8 filings for the US bankruptcy case, advisers in charge of liquidating the fund say that Zhu and Davies haven’t cooperated with them and that the founders’ whereabouts were unknown. On July 12, Zhu tweeted that a “good faith” effort to work with the liquidators “was met with baiting.” Zhu, Davies, and their lawyers did not respond to requests for comment.

The rise and fall of 3AC, as their Singapore-based fund was known, parallels the transformation of crypto. What started with speculation on a few well-known coins such as Bitcoin and Ether became an interdependent industry of tokens linked to other tokens, crypto companies acting like banks offering depositors double-digit yields, and lots of borrowing by traders looking to juice their returns. The rapid growth of this infrastructure helped push up crypto prices and 3AC’s fortunes; when prices turned this year, 3AC unraveled and may have even accelerated the decline. For all the complexity of the new crypto ecosystem—the “smart contracts,” reams of online white papers, and heady talk of decentralized finance—it still proved to be a giant wager on the simple idea that there would always be more buyers for digital coins and prices would mostly keep going up.

Crypto trading platforms including BlockFi and Blockchain.com have since disclosed exposures to 3AC. One—Canadian-listed Voyager Digital Ltd.—has gone under after 3AC defaulted on a loan worth more than $650 million. Its customers, which include a lot of ordinary investors, have had their crypto accounts frozen and are unlikely to get back all their assets. When the music stopped, it turned out that almost everyone had lent to, as many have called it, the Archegos Capital or even Long-Term Capital Management of crypto.

Crypto is vaunted for its transparency: Transactions are recorded on a public blockchain database, and many transactions are governed by open source software rules. But at the big money levels at which 3AC played, borrowing was largely a matter of relationships, not much different from the way typical hedge funds rely on the confidence of their bankers. 3AC borrowed from the biggest crypto lenders without disclosing much about its finances. Zhu’s and Davies’s online personas were populist billionaires who tweeted, “Only boomers trade stocks btw.” But among the pros, no one quite expected them to make such wild wagers. They turned out to be “degens,” crypto slang for degenerate gamblers, like everyone else, says one trading-firm executive who asked not to be named.

The duo’s background before crypto was conventional for finance. After graduating from the elite Massachusetts boarding school Phillips Academy and later Columbia, Zhu and Davies cut their teeth trading derivatives at Credit Suisse Group AG in Tokyo. In 2012 the two friends, then in their mid-20s. started their own hedge fund; it was a tiny operation that took advantage of price gaps across emerging-market currency derivatives, picking up the pennies in between while hoping those pennies would add up across lots of trades.

As digital assets started to take off in 2016, Zhu and Davies saw that the nascent crypto market was even more rife with the kind of pricing gaps they profited from in currency contracts. At one point in the heady days of 2021, Bitcoin futures were trading 50% above the “spot” price to buy the tokens themselves. So 3AC sold futures and bought spot, a classic Wall Street play that capitalizes on a temporary pricing disconnect.

Then came a bigger opportunity. The Grayscale Bitcoin Trust, or GBTC, allows people who can’t or don’t want to hold Bitcoins directly to instead buy shares in a fund that invests in them. For a while, GBTC was one of the few US-regulated crypto products, so it had the market to itself. It was so popular that its shares traded at a persistent premium to the of the Bitcoin it held.

Big investors such as hedge funds had a way to buy GBTC shares for less than what ordinary traders were paying, however. Grayscale allowed them to purchase shares directly by giving Bitcoin to the trust. An easy way to make money was to borrow Bitcoin, exchange it for shares, and then sell those shares at the premium price. At the time of its last filing at the end of 2020, 3AC’s was the largest holder of GBTC, with a position then worth $1 billion. But the strategy had a snag: The shares bought directly from Grayscale were locked up for six months at a time.

Starting in early 2021, that restriction became a problem. GBTC’s price slipped from a premium into a discount—a share was worth less than the Bitcoin backing it—as it faced stiffer competition from similar products. As the months went on, the discount got wider and wider. In early June, TPS Capital, a company that often brokered 3AC’s borrowing, tried to persuade other speculators to snatch up GBTC shares, according to two trading-firm executives who heard (and ignored) the pitch. They didn’t want to be named talking about private discussions. Timothy Chan, chief executive officer at TPS, says 3AC had come up with the trade and asked for referrals. At the time his firm wasn’t aware of any financial distress at 3AC, he says, and anyhow, to his knowledge, the pitch fell only on deaf ears.

The Grayscale trade was one of 3AC’s simpler plays. The crypto world, for a while, was filling up with strange new arbitrage opportunities that looked like ways for sophisticated investors to collect free money. This assessment seemed especially true in the buzzy corner dubbed decentralized finance. DeFi aspired to build a replica of Wall Street—with all of its deposit-taking, trading, borrowing, and insuring functions, but less of its regulation—on the blockchain.

To change the world, DeFi startups needed to get people to entrust their crypto tokens to them. At a time when savings rates on bank accounts were almost zero and safe bonds were earning less than 2%, DeFi platforms were boasting double-digit yields for depositors, generated in various ways. Like many others, 3AC both borrowed and deposited.

As crypto prices plummeted, the comedown for DeFi was brutal. The hottest fad in early 2022 was the Anchor Protocol, which offered a 20% rate. But to earn that, you had to hold TerraUSD, a token that a brash crypto influencer called Do Kwon conjured, which in turn was linked to another coin he created called Luna. The whole system counted on Luna being worth something, which in the best of times assumed a future in which everyone used the technology Kwon developed to trade tokens and digital art.

That future didn’t come soon enough for 3AC or the vast ranks of other “Lunatics.” Not only was 3AC earning yields on Anchor, but it had also invested $200 million in Luna in February, Davies told the . At their height, Luna and TerraUSD were worth a combined $60 billion. But when the price of Luna collapsed, it all vaporized.

After that, jitters spread through crypto. 3AC was also investing Ether on a platform called Lido Finance, trying to generate yield from something called staking. In a nutshell, Ether tokens are needed in a technical process for verifying blockchain transactions. If you agree to lock up your tokens for a long period of time to support this activity, you can earn more Ether as a reward in the future. Lido’s innovation is that when its depositors’ Ether is locked up, they get another tradable token called stETH. For most of this year, stETH has traded at par with ETH, but after Luna crashed, the price slid to as low as a 7% discount as traders rushed for the exit.

3AC was one of them. On June 14, it pulled more than 80,000 stETH (more than $84 million) from the DeFi lending protocol Aave in only four transactions and started converting them back to Ether at a lower price, according to data provider Nansen. It was a classic financial crisis dynamic: As soon as prices fell far enough, people became desperate to get out, even at a loss, helping send prices still lower. “What we’re seeing over this period is all of that bubbling up and feeding itself,” says David Fauchier, a crypto fund manager at Nickel Digital. “That’s what happened in 2008. This is your very typical liquidity crunch without a proactive central bank stepping in to do anything."

These trades—traceable on the public blockchain and described to by former employees who declined to be named because they aren’t authorized to speak publicly—could have been made by anyone who knew their way around crypto a bit. But Zhu and Davies had access to capital that your average Reddit crypto trader didn’t. They borrowed from large digital-asset lenders and wealthy hodl-ers, and reached brokerage agreements with JPMorgan Chase & Co. and Bank of America Corp. They had access to the funds of some decentralized-finance projects, and one trading firm has accused them of using $1 million that it held with 3AC to meet margin calls.

3AC in at least one case may have resisted sharing detailed information with a lender. In text messages revealed by Hodlnaut, which offers crypto savings accounts, Davies, through TPS, in May was asking to borrow crypto coins without collateral. After the lender listed its requirements, TPS said 3AC didn’t disclose an audited balance sheet and offered a statement of its net asset s instead. The would be self-declared and not contain a breakdown of its investments. Hodlnaut said it walked away.

“What surprises me most about 3AC’s collapse was how they were able to amass so much leverage in the first place,” says Ryan Watkins, co-founder of crypto hedge fund Pangea Fund Management. “It was the lack of transparency that both enabled 3AC to borrow so much money as well as cause panic across the industry, as no one knows who was exposed and how badly they were.”

In retrospect, 3AC has always been a mystery. The fund itself is in the British Virgin Islands, and the company is licensed in Singapore to manage other people’s money. But Zhu and Davies have always maintained that the pot—all $3 billion of it, from what Davies told the is entirely theirs.

To further complicate things, 3AC also housed two funds as sub-share classes: DeFiance Capital for DeFi investments and Starry Night Capital for digital art. DeFiance has external investors, and its founder has insisted it operates independently, but that structure has put its future in doubt. It’s now considering its legal options in light of 3AC’s bankruptcy, according to a person familiar with the matter who declined to be identified discussing an ongoing issue. 3AC’s relationship with TPS has also come under scrutiny. In the industry, TPS is known as 3AC’s “over-the-counter” desk, though it’s a separate company in which Zhu and Davies hold a stake. Last week it released a statement saying that even though the two companies have referred business to each other and it has coordinated loans for 3AC, its operations are distinct.

On June 30, the Monetary Authority of Singapore reprimanded 3AC, without a fine or other sanction, for providing false information and exceeding the limit on its assets under management. Before the debacle, 3AC was seeking to move to Dubai, which has welcomed the crypto industry. Only two months ago, Zhu and Davies were meeting with some of the world’s largest venture capital firms and sovereign wealth funds at Sequoia Capital’s conference in neighboring Abu Dhabi, according to people with knowledge of the matter. Some of the people say the pair had set up shop in a Dubai office building, although the signage has since been removed and a spokesperson for the complex says they have no office there. 

As 3AC began to unravel, Zhu and Davies met with executives from several crypto exchanges to discuss the possibility of a bailout—with no luck, say people with direct knowledge. The crypto market is now going through a classic downturn in the credit cycle, much like the global economy. Unlike the real world, where loans are made to start businesses or buy homes, the demand for crypto leverage comes almost exclusively from speculators. “One thing we all realized is crypto is much more correlated with outside worlds than it used to be,” says Evgeny Gaevoy, the founder of Wintermute, one of the largest crypto market makers. “A lot of centralized entities like Three Arrows—they further exacerbate this exuberance of the cycle. They make all the numbers going up much higher than they should have been.”

Davies handled most of 3AC’s external calls, and Zhu was an ideas man, ex-employees say. Zhu’s big thesis was the “supercycle,” an extended price rally fueled by a technological revolution that will build a decentralized internet on the blockchain. Zhu likes to evoke sweeping history from imperial China to the Dark Ages to paint a long path toward crypto dominance. Even in private dinners and group chats, the two men have argued their bullish convictions, acquaintances say. In May, Davies was still talking about buying Bitcoin and Ether on margin, says a person who met him then but wasn’t authorized to speak to media.

“I was just surprised at how sincerely they seemed to think this despite that they’re formerly forex traders,” says Haseeb Qureshi, managing partner at Dragonfly Capital, a venture fund. Currency traders, after all, should be accustomed to prices swinging both ways. “They might have been stupid and misguided, especially in a market like this that’s buckling under macro pressures,” he says. “But they really believe this stuff, and you can see it in their books, right? You wouldn’t have traded this way, if you didn’t believe that this was true."

In the cryptopian vision of a perfect future, one without centralized exchanges, something like 3AC’s collapse isn’t supposed to happen. People could lose money, sure, but everyone’s holdings would be visible on the blockchain. Reputations would be irrelevant. Losing bets that fall beneath a lender’s collateral requirements would be liquidated without mercy, and no one would be waiting around for a margin call. Thanks to these dynamics, Aave, one of the largest lending protocols, has survived just fine, the argument goes.

But that future seems far away. The crypto financial crisis of 2022 turned out to be like every other financial crisis: a case of too much trust as asset prices rose, followed by a sudden collapse of confidence. In May, Zhu called his “supercycle” theory “regrettably wrong.” On June 15, he tweeted that he and Davies were “fully committed to working this out.” By then, Zhu had quietly removed hashtags of protocols such as Luna from his Twitter bio. His header photo remains: three arrows pointing up, accompanied by the words “up only.”

More stories like this are available on bloomberg.com

©2022 Bloomberg L.P.