Solver auctions: inside the $23.3 billion market that settles your swap

01 · What happens after you press swap

How an intent becomes a trade

On Uniswap, and on most other DEXs, a swap goes into a pool. The pool holds both tokens, the curve sets the price, and the trade lands in the next block. While the transaction sits in the public mempool, anyone can see it. A sandwich bot can buy just ahead of you and sell just behind you, and you eat the slippage.

CoW Swap, a DEX, never sends the order to a pool. You sign a short message that says what you will sell and the least you will take back. That message is an intent. It stays off-chain. A group of firms, the solvers, compete to fill it: each one computes a route, the best bid wins, and the winner packs the order into a batch, pays the gas, and settles the trade on-chain.

On Ethereum about 30 of those firms bid every few seconds. The same contract, at the same address, now runs the same auction on BNB Chain, Arbitrum, Base and Polygon.

Traders use it because the fill is usually close to the pool price, and because an order that never sat in the mempool cannot be sandwiched. Solvers make money on the spread between the limit you signed and the price they actually got, and CoW pays them a COW token reward on top of that.

Bitquery indexes every call and event on Ethereum, BNB Chain, Arbitrum, Base and Polygon. We pulled the 2026 record for CoW Protocol and its three rivals, UniswapX, Bebop and Relay, to see how big this market is and who takes the pay.

02 · Five chains, one market

From June, BNB Chain settled more batches than Ethereum

CoW uses one settlement contract, at the same address, on all five chains we measured. What changed this year is where the traffic went. Ethereum used to be the busy venue. From June, BNB Chain settled more batches each month than Ethereum. July was the peak, more than two and a half times Ethereum's best month.

The July spike on BNB Chain came from wallets spread across the chain rather than a few addresses recycling the same flow. Base, which we can only read from June because of a gap in the decoded event record, arrived already at Arbitrum's level. Polygon barely moved, and its ten biggest solvers took nearly all of the batches.

03 · The dollars stay on Ethereum

Why the volume chart and the trade table disagree

The monthly chart is why the trade count and the dollar chart disagree. From June, BNB Chain settled more batches than Ethereum. The dollars did not follow. Ethereum's orders are large enough that it still accounts for almost all of the value.

In the first two weeks of September, among Ethereum orders in the major tokens we can price, the median is $734: line them up from smallest to largest, and the one sitting in the middle is that size. Half of that sample is at or below $734. A small number of very large orders pull the mean about 37 times higher, which is why the average is a misleading way to describe the typical trade. On BSC the typical ticket is a few hundred dollars.

The reason the cheaper chains absorbed the trade count is gas. A solver that wins an Ethereum batch pays a little over a dollar to settle it. The same job on Arbitrum costs about two cents, and on BSC even less. A dollar of gas barely registers on a $734 Ethereum order. It is a real cut of a few-hundred-dollar BSC swap, which is why the small tickets moved to the cheap chains.

Cheap gas also changes the shape of a batch. CoW's original idea was that several orders could cross in one transaction, so traders would take liquidity from each other instead of from a pool. On Ethereum and BSC a settlement still holds about one order, which means the auction is usually filling a single swap. On Arbitrum a batch holds about one and a quarter. Matching many orders against each other in one transaction is still rare on every chain we measured.

Even on Ethereum, where the dollars are, CoW is a thin slice of the whole market. Mainnet DEX volume in the same window is roughly $280 billion. CoW's share of that is something like seven or eight cents of each dollar.

04 · The solvers

Who wins the auctions

A solver is the address CoW writes into the settlement event as the winner of that batch. Public dashboards name most of the teams behind those addresses. In the first half of this month, 32 solver addresses won at least one auction on Ethereum.

A year-to-date board hides how fast the ranking moves. BRRRolver is tenth for the year. The top five on Ethereum still win under half of all auctions, which is the most open of the five venues. On BSC the top five take about three quarters, on Arbitrum about seven in ten, and on Base about two thirds.

A Base solver has no public name. It does not show up in CoW's own solver lists.

05 · What the job pays

36.8 million COW, and half of it to one team

When a solver wins, CoW pays it out of a rewards safe on Ethereum. Those transfers are public, so anyone can add them up.

The dollar values depend on the COW price, which we took as a median of two on-chain trades, so treat them as rough. They are also only the visible part of what a solver earns. If the solver fills you at a better price than your limit, it can keep some of the difference inside the batch. That surplus never shows up as a transfer. Adding the COW and the ETH gas refunds, the safe paid out about $3.6 million this year. Solvers burned about $1.0 million in gas. The refunds nearly cover the Ethereum gas bill.

The fill itself is usually close to the public market. Failed settle calls are rare on mainnet, about one in 500, and more common on Base, about one in 65. In a three-day sample in September, the median WETH-to-USDC order on CoW priced within five basis points of the same-block Uniswap median. For the trader, that means a Uniswap-like price and an order that never sat in the mempool, which is the sandwich protection.

06 · Who uses it

A month of Ethereum wallets, sorted by how often they trade

Most wallets that used CoW on Ethereum in the 30 days to mid-September placed a single order and did not come back. A thin slice of wallets, under one percent, placed more than half of the orders: bots, solvers testing routes, and desks that swap all day.

BNB Chain is less dominated by a few wallets. In July, 67,000 wallets traded and the ten busiest took a fifth of the orders.

About 7% of June's CoW orders on BNB Chain were sales of TUFT, a token tied to a pig-butchering scam. Victims had been told to swap it back to USDT, and CoW was the venue that handled those exits: about 24,000 orders this year, 3% of the chain's flow, which we traced with money-flow tools.

07 · The rival that left the chain

1inch Fusion's settlement contract has not traded all year

CoW is one of several protocols that fill swaps in an auction instead of a pool. UniswapX, Bebop and Relay all do a version of this. Relay's fills are all settled by a single in-house solver, so there is no open competition on that venue.

1inch Fusion used to compete with CoW for Ethereum order flow. Its settlement contract recorded zero transactions on Ethereum and on BSC this year. We checked twice, on two replicas. The website and the docs are still up. On-chain, that auction has been quiet all year. We could not see where the flow went.

08 · How we measured it

What a settlement is, and a number we had to correct

The counts come from the settlement contract CoW deploys at the same address on all five chains. A settlement is its Settlement event, an order is its Trade event, a trader is the order's owner, and a solver is the address the event names as the winner. The window is January through mid-September. Every figure was run twice on separate replicas. Duplicate-event scans and cross-checks between the archive and realtime indexes sit behind the numbers.

The decoded event record on Base is missing for a year before June, so Base covers three and a half months. The raw chain record is complete, and a four-day fragment from May 2025 proves CoW was live on Base a year before our window opens.

The volume figures are lower bounds. We count the sell side in major tokens at prices from the dates traded, and Polygon's figure is a 30-day snapshot. The intent-share estimate uses an outside benchmark for mainnet DEX volume, because raw trade amounts in any index carry decoder outliers. Execution quality was checked against same-block Uniswap v3 medians for the busiest pair, a proxy rather than a tick-level audit.

One number failed audit and was fixed before print. Our first Base settlement count came back twenty-seven percent low because of a bad filter; the corrected read, verified three ways on two replicas, is what you see above. As an outside anchor, CoW DAO's own August recap reports $2.88 billion for the month across every chain it supports; our majors-only August read sits just under it, as a lower bound should. All of it was pulled with Bitquery's trade API and price feeds.