DeFi United raised $222M to rescue rsETH. Most of it hasn't cleared governance yet.
The rsETH bailout is a live test of whether protocol treasuries can act as a collective lender of last resort. Three governance votes are the critical path.
CryptoVibe Desk · defi · rseth · kelpedao
- →DeFi United committed 100,356 ETH ($222.76M as of April 26) to restore rsETH backing, with Arbitrum DAO, Mantle, and Aave DAO among contributors.
- →Roughly 85,765 ETH of that total is still pending governance votes from three DAOs, meaning the $222.76M figure overstates the actual available backstop right now.
- →The rescue's headline number only materializes if Arbitrum and Mantle governance passes and KelpDAO reopens withdrawals, three independent events that DeFi United cannot control.
- rsETH → KelpDAO's token representing staked ETH in its restaking protocol, which lost its intended 1:1 backing with ETH after the April 18 incident.
- liquid restaking token → A token you receive when staking ETH through a restaking protocol, letting you use that token in other DeFi applications while your original ETH stays locked.
- DAO treasury → A pool of funds collectively owned by a protocol's governance token holders, which can only be spent after an on-chain vote approves the transfer.
- multisig → A shared wallet that requires a minimum number of designated signers to approve any transaction before it executes.
The DeFi United rescue effort isn't a charitable collection. It's the first live test of whether coordinated protocol treasuries can function as a lender of last resort inside DeFi, and the mechanics are messier than the headline number suggests.
As of April 26, DeFi United claims 100,356.559 ETH ($222.76M) raised across 83,003 wallets and 88,706 transfers, following the April 18 rsETH incident. The contributor list spans protocols and individuals: Arbitrum DAO, Mantle, Aave DAO, Stani Kulechov, Ether.Fi, Lido, BGD Labs, Golem Foundation, Emilio Frangella. But of that $222.76M, roughly 85,765 ETH comes from three line items each listed as pending governance vote: Arbitrum (30,765 ETH), Mantle (30,000 ETH), and Aave (25,000 ETH). The raised number is a commitment queue, not a settled balance.
The engineering problem here is coordination under external dependency. DeFi United's own site notes that recovery depends on third-party actions outside its control: KelpDAO reopening rsETH withdrawals, and the Arbitrum Security Council releasing frozen ETH. That's a dependency graph with two external nodes, either of which can stall the critical path. In a distributed system, you'd call this a liveness risk. The capital is there in principle; the settlement is not.
The tradeoff is legitimacy for speed. Individual contributors and foundations can transfer capital immediately. DAOs cannot, by design: each of the three pending contributors requires an on-chain vote that takes days to close. If KelpDAO's withdrawal window opens before the Arbitrum and Mantle votes clear, the effective backstop is materially smaller than $222.76M. DeFi United cannot compel the votes; it can only wait.
What's genuinely new is the model. DeFi hasn't seen a multi-protocol treasury recapitalization at this scale. The closest traditional-finance analogue is a syndicated liquidity facility during a credit event: multiple institutions pooling resources under coordination pressure to prevent a localized failure from going systemic. The difference is that the coordinating body here is a website, a multisig, and a set of governance forums operating simultaneously. It's a valid architecture. It's also fragile in specific, identifiable ways.
KelpDAO should publish a firm withdrawal-reopening date before the Arbitrum and Mantle governance votes go live, so delegates aren't committing treasury funds to a dependency that hasn't been confirmed.
Whether the Arbitrum (30,765 ETH) and Mantle (30,000 ETH) governance votes clear before KelpDAO publicly confirms a rsETH withdrawal reopening timeline, by May 10, 2026.
Primary links and supporting reads used by the desk for this story.
Forward this.
