THORChain Swaps Remain Paused on 12 Chains After Reported Refund Error

THORChain runs as a decentralized exchange where node-operated vaults swap native Bitcoin, Ethereum, Solana and Tron without wrapping them, a design it has operated since 2021 despite earlier vault exploits. Its security model assumes that once a deposit confirms on a source chain, it stays final. An Ethereum reorg broke that assumption: a transaction counted as settled was rolled back, and the refund logic processed it twice, exposing a gap between THORChain's cross-chain timing and the finality guarantees of the chains it bridges.

Node operators froze new swaps across 10 networks carrying Bitcoin, Ethereum, Solana and Tron, with the block extending to 12 chains in total. Traders and arbitrageurs lose the ability to route fresh trades through THORChain's pools until the halt lifts. Outbound signing kept running on most chains, so already-queued payments still settled, including transfers tied to a major exchange hack, meaning the pause slowed, rather than stopped, large flows already in motion through the protocol.

THORChain's premise is that swaps need no central custodian, only enough validators watching enough chains to agree on state. A reorg-triggered double refund shows that promise rests on finality assumptions that vary chain to chain, and that cross-chain infrastructure can fail in ways a single-chain protocol would not. The same mechanism that lets it move assets outside centralized exchanges also makes it harder for investigators to trace illicit funds once they are routed through it.

Resolution depends on whether node operators identify and patch the reorg-handling logic that produced the double refund before reopening the remaining blocked routes among the 12 chains. Until they vote to restore new swap intake, outbound signing will keep clearing existing orders but no fresh liquidity will move through the protocol. A full reopening across all chains, rather than a partial one limited to lower-risk networks, would signal operators consider the underlying vault code fixed.

The move was larger before the first headline than after. That is an observation, not an explanation: the market also moves on things nobody publishes.