622 Bitcoin. That is the sum at the center of a proposed class action against BitMEX. Not a theoretical loss from a smart contract exploit, but a demand for assets allegedly seized during forced liquidations. The complaint, filed in the Southern District of New York, accuses the exchange of operating an internal trading desk that traded against its users, freezing accounts during volatility, and executing unfair liquidations. Zero trust is not a policy; it is a geometry. Here, the geometry collapsed.
BitMEX was once the cathedral of crypto derivatives. It invented the perpetual swap, leveraged up to 100x, and captured the world's trading volume. By 2020, it faced CFTC charges for illegal off-exchange trading and failure to implement AML/KYC. Its founders stepped down and paid fines. Now, the platform is set to cease operations on September 23, 2026. The proposed class action is the final chapter of this saga. It seeks the return of 622 BTC—approximately $40 million at current prices—for a subclass of users who claim they were victimized by the exchange's internal trading desk and its liquidation engine. The code does not lie, but it often omits. BitMEX's internal logs omitted the extent of the conflict.
The core of the complaint: three counts of systematic failure.
First, the internal trading desk. BitMEX ran a proprietary trading operation that could see user positions, pain points, and liquidation thresholds. This is not a theory; the complaint cites internal communications where desk members discussed hedging against customer flows. From my audits of similar platforms, the internal trading desk structure is almost always a conflict of interest waiting to be exploited. TradFi exchanges like CME segregate proprietary desks with strict information barriers. BitMEX had no such barriers. The desk had access to the same database that held user margin and liquidation prices. That asymmetry is not a bug—it is a feature of the centralised trust model. Security is the absence of assumptions. BitMEX assumed internal trades were benign.
Second, forced liquidations. The plaintiffs allege that during the March 2020 crash and the May 2021 sell-off, BitMEX's liquidation engine failed to execute at fair market prices. Users claim their positions were liquidated at prices far below the oracle mark, generating excess profit for the internal desk. The mechanism is straightforward: when a liquidation cascade hits, the engine sells into a thin order book. If the exchange's own desk is the first to buy, it captures the spread. The insurance fund is supposed to cover losses, but here the fund's transparency was zero. The plaintiff's lawyer stated that "BitMEX's trading desk captured the liquidation profits that belonged to the users." Compiling the truth from fragmented logs. The only logs that prove this remain on BitMEX's servers.
Third, account freezes. During the same volatile periods, BitMEX allegedly disabled withdrawal functions for certain accounts, locking users out of their funds while liquidations occurred. The complaint frames this as a coordinated effort to prevent users from moving collateral to other platforms. In a downturn, the ability to withdraw is the only defense against a predatory liquidation engine. By freezing accounts, BitMEX removed that defense. The result: users who would have survived the volatility were forcibly closed.
Contrarian: What did BitMEX get right?
The bear case is not without merit. BitMEX was a pioneer. Its perpetual swap funding rate mechanism was a genuine innovation that stabilised premium and discount. Its insurance fund, though opaque, did survive several black swan events. The proposed class action faces legal hurdles: proving that the internal desk caused specific losses requires granular trade-by-trade analysis. BitMEX will argue that liquidations were automated and that the desk operated within its risk parameters. The exchange is already shutting down, so the practical impact may be limited to this claim. Bulls may argue that the suit is a relic of an older era, and modern exchanges like Bybit and OKX have improved compliance. But the core issue—whether a centralized exchange can ever be a fair counterparty—remains unanswered. The code is law only if you can read it. BitMEX's code remains closed.
Takeaway: The precedent matters more than the BTC.
This case is a stress test for centralized exchange accountability. If the court rules against BitMEX, it sets precedent that exchanges are fiduciaries with a duty to not trade against their customers. If the suit fails, the industry returns to caveat emptor. Either way, the 622 BTC is a down payment on a question that every CEX must answer: Are you a broker or a counterparty? The plaintiff's lawyer will push for discovery. If internal trading logs are unsealed, the entire industry will face a moment of truth. Will other exchanges open their liquidation logs, or will they wait for their own 622 BTC reckoning? The geometry of trust demands transparency. BitMEX had none.