CoW Protocol

• CoW Protocol executes trades through batch auctions rather than continuous order matching.

• Where two users want opposite sides of a trade, they settle directly against each other without touching a liquidity pool.

• COW is the governance token.

• Governance over protocol parameters, solver requirements and treasury.

• Solvers stake COW as a bond against poor execution.

Pros:

• Batch auctions give every trade in a batch the same clearing price, removing the ordering advantage that enables sandwich attacks.

• Direct matching between opposing trades saves the fees a pool would charge.

• Substantial volume, particularly among sophisticated and institutional users.

Cons:

• Batch settlement introduces a short delay compared with immediate execution.

• Depends on a competitive solver set; concentration would weaken execution quality.

• Governance token with no direct fee capture.

• Exposure to a trading protocol with a structural answer to front-running.

• Real volume from users who understand execution costs.

• Weigh the token's limited claim on the revenue the protocol generates.

• Uniform clearing prices within each batch, which removes the transaction ordering advantage that sandwich attacks depend on.

• On transparent blockchains, anyone can see a pending trade and position ahead of it, extracting value from the trader.

• Settling a batch at one price makes ordering within the batch worthless.

• Emerged from Gnosis, founded by Martin Köppelmann and Stefan George.

• Operates as an independent protocol with DAO governance.

• No public ICO. COW was distributed via airdrop and a community sale in 2022.

• Solver concentration would degrade execution quality.

• Batch delay is unsuitable for latency-sensitive trading.

• Token captures no protocol fees directly.

• Competition from aggregators and intent-based systems.

CoW Protocol's papers begin from the observation that a coincidence of wants — one user selling exactly what another is buying — should be the cheapest possible trade, requiring no liquidity pool and paying no pool fee, yet conventional exchanges never find these matches because they process orders one at a time. Collecting orders into batches and clearing them at a single uniform price makes such matches discoverable, and simultaneously destroys the value of ordering within the batch, which is what front-running extracts. Solvers compete to propose the settlement that gives traders the best aggregate outcome. The design's dependency is stated plainly: the guarantee is only as strong as the competition among solvers.