Trader Took $1.28M From Papertrade by Moving Hyperliquid's ETH Price

A single wallet took about $1.28 million out of Papertrade during the venue's first evening of trading, by moving the price Papertrade quotes against. The exchange is a perpetual-futures market on HyperEVM, Hyperliquid's smart contract layer, and it opened at 10 a.m. New York time on Saturday with leverage up to 1,000x. Each trade there settles at the midpoint of Hyperliquid's best bid and offer. So the wallet opened large Papertrade positions and then traded ETH on Hyperliquid to shift that midpoint. The Block reported the total on Sunday evening in New York, and TokenPost filed the same figure later that night.
Three dollars was enough
The arithmetic is the whole story. At 1,000x leverage, a move of 0.1% in the reference price is the entire margin on a position. Between 6:01 and 6:02 p.m. Eastern the wallet bought roughly 6,060 ETH on Hyperliquid and the price rose about $3, or 0.12%. A minute later it sold about 5,210 ETH and the price gave back about $3.50. At 6:06 p.m. it withdrew $1.54 million. Across Saturday the account's withdrawals came to about $2.48 million against $1.2 million deposited, which is where the difference of $1.28 million comes from.
The second leg went the other way
At 7:13 p.m. the wallet opened 30 ETH shorts at 660x leverage, a combined notional of about $297 million, then sold about 6,590 ETH over three minutes and took the reference price down to $2,504.40. Those shorts did not close. They sat open for more than three hours and were liquidated at 10:31 p.m. at $2,509.95, costing roughly $450,000 and cutting into the earlier gain. The liquidation minted about 17.6 million PAPER, the token Papertrade pays out to losing traders, to the same wallet. Why the positions stayed open is not established. An X user who published the wallet address that night suggested Papertrade may have blocked it from closing them, a reading that is one account's and that the exchange has not addressed.
What the exchange says, and what nobody has confirmed
Papertrade posted on X on Sunday. It described the overnight attempts on its bitcoin and ether markets in terms that put the fault outside its own design.
The attempts were "not an oversight in the mechanism design we spent six months working on," Papertrade said.
The cause it gives is its own risk settings — launch-period open-interest caps left loose to absorb demand. TokenPost renders the same line as caps "set unusually high to handle demand during launch". Papertrade said it has begun tightening those caps on-chain and at the relayer level, and that its launch queue made the timing of opens and closes hard to predict. It also said the profiting trader later returned almost all of the gains through new accounts, and that some users agreed to reverse manipulated trades in exchange for keeping the PAPER they had earned. The Block could not verify either claim, and Papertrade's own records still showed withdrawals exceeding deposits by about $1.28 million. TokenPost hit the same wall. Neither account confirms the return.
The dial that is being turned
Open interest is the measure Papertrade now says it is constraining, and on Hyperliquid itself that figure has been the one to watch, with open interest reaching $18 billion in September. Prices on that venue family have moved in ways the underlying did not before, including the July session when an SK Hynix perpetual fell 18.7%. What is different here is that nothing broke. The pricing worked as documented and the leverage was as advertised. The loss landed on a shared pool that is funded by other traders' losses, and Papertrade has published no post-incident accounting of what the pool absorbed. The caps it describes tightening carry no figures either. Until both are published, the only measure of the damage is the exchange's own interface.
Read also: Bonzo Lend Loses $9.05M to a Flaw in a Price Oracle