CoW DAO

CoW DAO is the community that governs CoW Protocol, a trading protocol built around intents rather than direct swaps. It is one of the cleanest examples of GnosisDAO's "on-chain holding company" pattern: CoW (originally "CowSwap") began as an internal Gnosis experiment, then graduated into a standalone project with its own token and treasury via GIP-13 (Phase 2) and a $23M round in January 2022. The name is a pun on Coincidence of Wants – when two traders want opposite sides of the same trade, they can be matched directly against each other instead of both paying an AMM.

Intents and batch auctions

Most DEXs execute one swap at a time against a liquidity pool, which exposes every trade to MEV – front-running, sandwiching, and priority-gas auctions that quietly tax the trader. CoW Protocol restructures this. A user signs an intent ("I want at least X of token B for my token A") rather than a transaction, and intents are collected into discrete batch auctions. Independent third parties called solvers then compete to settle each batch: they net off any coincidences of wants peer-to-peer, route the remainder across on-chain liquidity, and the winning solver is the one that returns the most surplus to users.

Every trade in a batch clears at the same uniform price, which removes the ordering advantage MEV bots rely on. The protocol also ships CoW AMM, a liquidity-provision design that protects LPs from loss-versus-rebalancing. It also co-created MEV Blocker, an RPC endpoint that shields ordinary wallets from sandwich attacks, but no longer runs it: the RPC was acquired on 23 January 2026 by Special Mechanisms Group, the blockchain-microstructure division of Consensys, with the DAO supporting the transition and keeping MEV protection in CoW Protocol and CoW Swap. The same intent-and-solver shape appears in the 1inch DAO's Fusion mode, where competing resolvers settle gasless orders through Dutch auctions.

Governance: COW, CIPs, and Snapshot

CoW DAO controls the protocol treasury – 44.4% of the one-billion fixed COW supply – plus solver approvals, fee parameters, and grants. Governance follows the same optimistic pattern its parent GnosisDAO pioneered: ideas are debated as CoW Improvement Proposals (CIPs) on the CoWmunity forum, then ratified by a gas-free Snapshot vote – off-chain signalling that costs holders nothing. Team and investor allocations were issued as non-transferable vCOW that vests linearly over four years and converts 1:1 to COW, aligning insiders to the long term. Because 10% of supply sits with GnosisDAO, the two DAOs remain economically intertwined – a recurring feature of the spun-out-subsidiary model.

The cow.fi domain hijack and CIP-86 (2026)

On 14 April 2026 CoW Protocol's domain registrar (Gandi) was compromised via social engineering, letting attackers hijack DNS for the cow.fi domain for roughly 4.5 hours and serve a phishing UI that drained wallets of an estimated ~$1.2M. The protocol's smart contracts were never breached – the failure was entirely at the Web2 edge. What happened next is the governance story worth studying: rather than disclaim responsibility for an off-chain attack, the DAO passed CIP-86, a discretionary grants program reimbursing eligible phished users up to 100% of losses from the treasury. The DAO was explicit that the payments were voluntary "goodwill" grants, not an admission of legal liability – a deliberate framing that let a treasury-funded reimbursement pass on Snapshot without setting a liability precedent. It is a clean example of a DAO treating its treasury as a stakeholder-protection backstop.

How Caper approaches this

CoW DAO's CIP-86 vote is a reminder of what a treasury is for – a shared pool that can be deployed to protect the people who make the protocol valuable. But reaching it required a bespoke proposal, a Snapshot quorum, and a discretionary grants process stood up after the fact. Caper bakes the members-first claim on the treasury into the protocol itself: every caper lets a member who has voted exit at any time for their fair share of the treasury, with no proposal or quorum, and the same earned vote weight that sets their say governs what they take – the influence you hold and the value you can walk away with go together. Where CoW DAO had to vote its treasury into acting for members, Caper makes the members' standing claim a protocol default. (Caper does not attempt CoW's intent-based trade execution – the parallel is purely about treasury governance.)