Chainflip

Chainflip’s answer to cross-chain trading was to leave each asset in its own world. A user sends BTC to a Bitcoin vault; validators witness the deposit; liquidity providers compete with just-in-time quotes on State Chain; enough validators jointly sign the payout to meet the required threshold, releasing native ETH, SOL or a stablecoin from the destination vault. No FLIP-denominated claim on the incoming coin is created.

That choice removes a wrapper issuer but creates a choreography problem. The vaults, brokers, State Chain and destination networks must agree on which deposit belongs to which recipient. FLIP sits behind that choreography as validator collateral: it helps select the authority set and exposes bonds and delegations to slashing when operators fail.

FLIP became transferable at noon UTC on 23 November 2023, before unrestricted swapping. On 13 December Chainflip opened a production pre-release after more than three years of development and two years of testnets. The team said its JIT AMM had never run in the wild and that testnet could not teach market makers how real capital would behave.

So the first door was deliberately narrow. Frontend and backend swaps used asset-specific maxima worth about $1,000 per deposit, aggregation and cross-chain messaging stayed off, and Chainflip warned that an amount sent above a hard-coded limit could be swallowed. Validators, brokers, wallets, liquidity providers and the protocol team could now tune their separate systems with real money while limiting what a bad handoff could cost.

Chainflip reported more than $1 billion in swaps during 2024. Boost accelerated deposits, DCA split large orders over time, and native SOL plus Arbitrum widened the routes. The figures are the project’s account, but they mark how quickly the restricted launch became a functioning exchange used through outside wallets and aggregators.

On 4 February 2026 the same machinery expanded into public lending for BTC, ETH, SOL, USDC and USDT. Chainflip’s lending design uses its built-in exchange to unwind collateral rather than asking an external liquidator to take the whole position at once. The product made the vault network serve ongoing credit positions, not just one-time swaps.

The expansion also increased the number of boundaries that had to agree. A user could see one interface while a broker created channels, State Chain indexed events and several chains executed deposits or refunds. The June 2026 failure appeared at one of those less visible boundaries, not in the advertised native-asset route itself.

On 11 June, a broker still running version 2.1.13 relied on a transaction index that no longer reliably identified an extrinsic. Two deposit-channel creations landed in the same block and the broker paired one channel ID with the other channel’s address. A user followed the displayed instructions; Chainflip processed the deposit and paid the destination attached to that different channel.

The fix had already arrived in version 2.1.16 and identified extrinsics by cryptographic hash. Chainflip upgraded the affected broker and told all brokers to move to version 2.1.20. Its report says investigators were trying to contact the parties, but does not say whether the user recovered the funds. A stale integration had defeated a swap even though vault signatures and AMM settlement did what their inputs told them to do.

In August, Chainflip detected an attempted exploit in Ethereum messaging and refund logic. It paused Ethereum deposits and quotes, paused Arbitrum and Tron as precautions, held affected swaps, and said the completed fix took about 24 hours from first report. The project reported no user loss and said held swaps were being processed and sent automatically or would be refunded without user action. The pause protected funds by making the supposedly seamless route visibly unavailable.

In a separate September event, Chainflip accused three validators of sending malformed key-generation messages. It paused authority rotations while operators upgraded. Proposal 007 asked to move those validators' FLIP to an on-chain reserve and exclude them from future key ceremonies; passage and execution were not yet established. Chainflip set out to remove the wrapper from cross-chain trade. Its operating history shows where responsibility went: brokers identify deposits, validators share vault keys, Safe Mode closes dangerous routes, and FLIP puts collateral behind the people running them. A native swap therefore depends on those handoffs continuing to identify the same deposit, recipient and network state.