Papertrade tightened trading-risk controls after a wallet used relatively small movements in Hyperliquid's Ether market to amplify gains on Papertrade's newly launched synthetic perpetuals exchange. Public records reviewed by The Block showed the wallet withdrew about $2.48 million after depositing roughly $1.2 million, leaving a net difference of about $1.28 million. The episode is a live test of how extreme leverage and externally sourced prices can interact during a protocol launch.

Small ETH moves became large leveraged outcomes

Papertrade offers synthetic BTC and ETH positions with leverage of up to 1,000 times and uses the midpoint between Hyperliquid's highest bid and lowest ask as a reference price. According to The Block, wallet 0x2025…840e repeatedly opened large Papertrade positions, then bought or sold thousands of ETH on Hyperliquid. One sequence between 6:01 p.m. and 6:03 p.m. Eastern time moved the midpoint only about $3 to $3.50, or roughly one-tenth of one percent. At 1,000-times leverage, however, a 0.1% move can equal the trader's entire posted margin.

Related reporting: NEAR Intents Says Full $3.8 Million Was Returned After Cross-Chain Exploit

Onchain activity shows the cost of moving the reference market

The Block calculated that the wallet traded about $231 million of ETH on Hyperliquid and paid around $52,000 in fees. Papertrade's portfolio interface showed approximately $1.28 million in realized gains on $3.18 billion of trading volume. Those figures illustrate that the tactic was not costless: the trader had to create market impact on Hyperliquid while carrying highly leveraged exposure on Papertrade. They do not, by themselves, establish the trader's intent or prove a violation of any rule.

Papertrade says launch caps were too permissive

In an official October 11 statement, Papertrade described the activity as overnight manipulation attempts and said the risk was possible because open-interest caps were extremely permissive during launch congestion. The team said it began tightening those caps both in its onchain contracts and at the relayer layer as the initial queue cleared. Papertrade framed the goal as making attempted price manipulation economically unattractive while preserving neutral treatment between the house liquidity pool and traders.

Later losses complicate the final tally

The same wallet later opened 30 ETH short positions at 660-times leverage, representing about $297 million of notional exposure, and was liquidated after more than three hours for a loss of roughly $450,000, The Block reported. Papertrade also said the user returned almost all of the earlier profit through new accounts, while some other users agreed to reverse trades in exchange for PAPER rewards. The Block said it could not independently verify those statements, and the public records it reviewed still showed withdrawals exceeding deposits by about $1.28 million.

The house pool concentrates counterparty risk

Unlike a conventional order book in which traders face one another, Papertrade's protocol-owned pool acts as the counterparty and pays winning positions from trader losses and available liquidity. The Defiant reported that its dashboard later showed a pool balance near $3.14 million, while other snapshots varied as positions and withdrawals changed. For depositors, that structure makes open-interest limits, liquidation design and the resilience of the reference price central risk controls rather than secondary settings.

What remains unresolved

Papertrade has confirmed that it changed exposure limits, but it has not published a complete post-incident accounting that reconciles every wallet, reversal and pool movement. The available evidence supports describing a pricing and risk-control incident, not a confirmed theft or protocol exploit. The next useful disclosure would be a wallet-level reconciliation and a clear explanation of how the revised caps behave during volatile or thin market conditions. Until then, the final economic loss to the pool and the effectiveness of the new controls remain uncertain.

Sources

AI-generated editorial image; not a photograph of the reported event. Prepared with AI assistance and source verification.