Anna George

Anna George is the co-founder and CEO of CoW Protocol — the decentralised exchange infrastructure project that reimagined how trades are settled on Ethereum by introducing batch auctions and Coincidence of Wants (CoW) matching as an alternative to the fragmented, MEV-vulnerable AMM model that dominates most DEX trading. George came from a background in product management and operations within the Ethereum ecosystem, having worked in roles that gave her direct experience with the UX and execution quality problems facing DeFi users before founding CoW Protocol.

CoW Protocol originated within Gnosis — the prediction market and Ethereum infrastructure company founded by Martin Köppelmann — as Gnosis Protocol v2, a redesigned settlement layer intended to address the fundamental execution quality problems inherent in constant-function AMM trading. George was central to the product design, go-to-market, and eventual spin-out of the protocol into an independent project operating under its own DAO governance structure.

The core innovation of CoW Protocol is deceptively simple: rather than routing every trade immediately to on-chain liquidity pools (where it is visible to MEV bots that can frontrun or sandwich it), CoW Protocol collects orders into batches and first attempts to match them peer-to-peer — finding Coincidences of Wants where one user's sell order directly satisfies another user's buy order at the same or better price, with no AMM fees or slippage.

Only the volume that cannot be matched peer-to-peer is settled against on-chain liquidity. This two-step process — CoW matching first, then AMM routing — gives users better prices on average than direct AMM trading, eliminates frontrunning on the peer-to-peer matched portion, and dramatically reduces the gas costs and slippage associated with large trades. The batch auction structure means that even the on-chain settlement component happens in a single settlement transaction per batch, rather than individually, spreading gas costs across multiple traders.

The settlement is handled by a competitive network of 'solvers' — professional market makers and algorithm designers who compete to find the best execution path for each batch of orders and are rewarded based on the surplus they generate for traders versus a reference price. This solver competition internalises the search for best execution rather than leaving it to individual users who lack the tools or expertise to optimise their own trades.

Maximal Extractable Value (MEV) — the profit that can be extracted by miners, validators, or sophisticated bots by reordering, inserting, or censoring transactions within a block — has cost Ethereum DeFi users hundreds of millions of dollars. The most common forms, sandwich attacks (placing buys before and sells after a large pending trade) and frontrunning (racing to execute a trade before a detected profitable transaction), exploit the transparency of the Ethereum mempool where pending transactions are visible before they are included in a block.

CoW Protocol's architecture is inherently resistant to these attacks. Because users sign orders off-chain (not broadcasting them to the public mempool) and settle in batches via solvers, there is no publicly visible pending transaction for MEV bots to exploit. The batch auction model also makes sandwich attacks structurally impossible — the uniform clearing price mechanism means all trades in a batch execute at the same price, removing the price differential that sandwich attacks exploit.

George has been an articulate public advocate for MEV protection as a user rights issue — arguing that DeFi platforms have an obligation to protect users from value extraction by sophisticated actors, and that the transparency of blockchain settlement should not come at the cost of user execution quality.

Building on the solver competition infrastructure, CoW Protocol launched CoW AMM — an automated market maker that captures MEV for liquidity providers rather than allowing it to be extracted by external bots. Traditional AMMs (Uniswap, Curve) are routinely arbitraged against each other, with the arbitrage profit flowing to the arbitrageurs rather than the LPs whose liquidity made the trade possible. CoW AMM's design routes rebalancing through the solver network, capturing arbitrage value for LPs and reducing impermanent loss.

Under George's leadership, CoW Protocol has grown into one of the most significant DEX aggregation and settlement infrastructure layers on Ethereum, processing billions in monthly trading volume and becoming a key component of the 'intent-based' trading paradigm — where users express what outcome they want (best execution on a given trade) rather than specifying how to achieve it, leaving the how to competing solvers. This model has become influential beyond CoW Protocol itself, inspiring a broader shift in DeFi UX design toward intent-centric architectures.