On-chain investigationOptimismDEX tradingJan to Sep 2026

Is anyone trading on Optimism? Bots are. People trade under $3 million of its $20 million per day.

Between January 1 and September 20, 2026, the exchanges on Optimism cleared $5.89 billion of swaps. We rebuilt all 84.5 million of them, then sorted the wallets behind them into bots and people.

At a glance
Optimism is a layer 2 for Ethereum. Its main exchange is Velodrome. We counted every DEX swap on it this year. The money mostly stayed. Volume still runs at about $20 million per day, and it fell less than on Base or Arbitrum. The crowd did not stay. About 163,000 wallets traded in January and about 23,000 in August. Since July, bots make 86% of the volume by our rules. Three-quarters of this year's trading is ETH against a stablecoin, or one stablecoin against another.
$5.89B
DEX volume on Optimism, January 1 to September 20, 2026
86%
Share of volume made by bots since July, by our rules
$2.65M
Traded per day by people since July
-86%
Wallets that traded, January against August
Key findingsEach line has its own share link
  1. -30%

    Optimism's DEX volume per day, January to September. On the same count Base fell 41% and Arbitrum 68%.

  2. 86%

    Of Optimism's DEX volume since July came from bots, by our rules. People trade under $3 million per day.

  3. 163k → 23k

    Wallets that traded on Optimism's DEXs in January, and in August.

  4. $722M

    Of first-quarter swaps came from smart accounts churning USDC and USDT, 72% of it through OKX's DEX router. It faded after March.

  5. 75%

    Of all volume is ETH against a stablecoin, or one stablecoin against another. Memecoins, new tokens and OP ecosystem tokens add up to 2.5%.

  6. $268M

    Traded through one Uniswap V4 pool whose swaps log zero amounts. DEX trackers that read swap events miss it, ours included.

01 · A quiet SaturdayOptimism still trades. The question is who.

On Saturday, August 15, the exchanges on Optimism handled $4.9 million of trades. It was the slowest day of the year. On January 21, the day after a sharp swing in the ETH price, they had handled more than 30 times as much.

Optimism is a layer-2 network for Ethereum. It runs its own chain, where trading is cheap, and posts its results back to Ethereum. It has its own coin, OP. Its main exchange is Velodrome, a DEX built for Optimism. Coinbase's Base runs on the same code, the OP Stack.

So is anyone still trading on Optimism? We rebuilt every DEX swap on the chain this year from Bitquery's raw on-chain data. Then we sorted the wallets behind them into bots and people.

Most of the money stayed. Optimism still trades about $20 million per day, and it lost less of it this year than Base or Arbitrum. The people left. About 154,000 people traded on its DEXs in January, nearly two in three of them through bridge apps such as Relay's. Since July it has been under 20,000 per month, and bots make 86% of the volume by our count.

One pool even hides its trades from DEX trackers. About $268 million went through it between late July and early September, and it shows up in nobody's volume, ours included.

02 · Next to Base and ArbitrumOptimism held up better than both

Start with the money. We put Optimism next to two bigger layer 2s, Base and Arbitrum. All three traded less in September than in January. It was a rough year for ETH, which lost almost half its price by June and then won back more than half of the drop.

DEX volume per day, January = 100Monthly average of daily volume, 2026 (September: 1 to 20)0255075100JanFebMarAprMayJunJulAugSepOptimism 70Base 59Arbitrum 32

We measured Base and Arbitrum from Bitquery's DEX data, priced the same way. From January to September, DEX volume per day fell 41% on Base and 68% on Arbitrum. On Optimism it fell 30%, and one burst of stablecoin swaps early in the year explains all of it. Leave that burst out and September matches January.

Size is another story. Base still trades about 33 times as much as Optimism, and Arbitrum about 7 times. In January, Arbitrum was 15 times bigger.

Per dayOptimismBaseArbitrum
Jan$35.1M$1,355M$526M
Feb$26.3M$995M$493M
Mar$26.4M$780M$351M
Apr$17.5M$786M$243M
May$14.5M$967M$161M
Jun$20.2M$1,163M$198M
Jul$17.9M$863M$132M
Aug$20.2M$744M$142M
Sep*$24.4M$803M$167M
*September 1 to 20. Base and Arbitrum from Bitquery's DEX data, with pools that faked trades removed.

03 · What gets tradedETH and dollars, mostly on Velodrome

That money buys very little variety. Two kinds of trade take three-quarters of it. Someone swaps ETH for a stablecoin, or one stablecoin for another.

What gets traded on OptimismShare of DEX volume by pair type, January 1 to September 20, 2026ETH ↔ stablecoin44.0%Stablecoin ↔ stablecoin30.7%Bitcoin pairs11.2%OP pairs6.2%ETH ↔ staked ETH2.9%Same asset (ETH/WETH, USDC/USDC.e)1.6%OP ecosystem tokens1.5%Long tail (memecoins, new tokens)1.0%Other majors (LINK, AAVE …)0.8%

Two pairs do most of the work. USDC against WETH is 38% of all trading, and USDC against USDT is 27%. Add bitcoin and OP and very little is left. Memecoins, new tokens and Optimism's own ecosystem tokens, such as VELO and SNX, come to 2.5%.

Velodrome runs the market. Its pools carry 58% of all trading, nearly all of it in Slipstream, its concentrated-liquidity pools. Uniswap comes second, with its V3 and V4 pools together near 29%. The ten biggest pools take 69%, and you can pull every one of their trades from Bitquery's DEX API.

DEXVolumeShare
Velodrome Slipstream$3,315.6M56.3%
Uniswap V3$1,140.1M19.4%
Uniswap V4$576.3M9.8%
WOOFi$368.9M6.3%
Solidly V3$281.6M4.8%
Velodrome V2$107.6M1.8%
30 others$99.9M1.7%

04 · The January churnSmart accounts swapped stablecoins in circles, then stopped

January looked like a busy chain. Volume ran at $35 million per day, and more wallets traded than in any other month.

Who made the volume, per dayAverage DEX volume per day by month, $ million, 2026 (September: 1 to 20)BotsPeopleSmart accounts (4337)Solvers, vaults, fillers01020304035.1Jan26.3Feb26.4Mar17.5Apr14.5May20.2Jun17.9Jul20.2Aug24.4Sep

A big part of it came from smart accounts. These are wallets run by a contract rather than a single key, built on the ERC-4337 standard. From January to March they made $722 million of swaps. Nearly all of those swaps, 94%, sat in five USDC/USDT pools. Three of them were Uniswap V4 pools with fees close to zero.

Most of that money went through one door. OKX's DEX router carried 72% of the smart-account dollars. For two-thirds of the operations, a paymaster picked up the gas. A paymaster is a contract that pays the fee so the user does not have to.

Swapping one dollar for another again and again, at almost no fee, with the gas paid, is the pattern of volume farming. We found nothing on-chain that names the campaign. It faded after March. By April, smart-account volume was down to 8% of March's level.

05 · After MarchThe dollars came back. The people did not.

From April, the number of wallets trading each month fell to a fifth of January's or less. It has stayed there.

Wallets that traded each monthDistinct wallets with at least one DEX swap, 2026 (September: 1 to 20)PeopleSmart accounts (4337)Bots and everyone else0k50k100k150k200k163kJan127kFeb117kMar32kApr30kMay24kJun22kJul23kAug20kSep

The dollars did something else. Leave out the smart-account churn and trading slid into May, then climbed back. On that count, September's $24.3 million per day is the best month of the year.

People's dollars shrank much less than their headcount. They traded about $4 million per day early in the year, and between $2 million and $4 million since April. Their number fell by almost nine in ten, from about 154,000 in January to under 20,000 per month since July.

Most of the people who left had come through bridge apps. In January, nearly two in three people trading on Optimism's DEXs arrived that way, mostly through Relay's app. Line those wallets up by what they traded, and the one in the middle, the median, moved less than 10 cents all quarter. One Relay contract alone drew about 117,000 wallets in the first quarter, and a few thousand in the third.

2026TotalNo smart
accounts
PeopleBots
Jan$35.1M$23.7M$4.3M68%
Feb$26.3M$21.3M$4.2M72%
Mar$26.4M$19.1M$2.3M70%
Apr$17.5M$16.9M$2.7M80%
May$14.5M$14.4M$2.5M80%
Jun$20.2M$20.1M$3.2M83%
Jul$17.9M$17.9M$2.2M86%
Aug$20.2M$20.1M$2.5M87%
Sep*$24.4M$24.3M$3.6M85%
Volume per day. "People" leaves out smart accounts. "Bots" is the share of all dollars made by bots, counting smart accounts that trade like bots, which the chart above shows as their own layer. *September 1 to 20.

So what is left is a smaller group of bigger players, and bots bring most of the money.

06 · Who is leftBots, and the biggest of them track Binance

We tag a wallet as a bot if it trades through a contract built for bots, sends more than 100 trades on a day it is active, or spends half its trades on arbitrage. By that rule, bots made 77.5% of the year's volume, and 86% since July. Tighten or loosen the rules and the share moves, but it never falls below two-thirds.

Very little of it is classic atomic arb, where one transaction buys in one pool, sells in another and ends with more of the same token. Atomic arb makes up 29% of all transactions but only 7% of the dollars. It is the same kind of trade we counted in our study of arbitrage across five chains.

The big money follows Binance. We took every June swap in Optimism's ETH/stablecoin pools. Then we checked which way ETH had moved on Binance in the minute before.

Trading behind Binance, June 2026
Optimism's ETH swaps followed Binance's last move far more often than chance.
69%
Of swap dollars went the same way as Binance's move in the minute before
50%
What random trading would give
48%
Went the same way as Binance's move in the minute after

This is CEX-DEX arbitrage. ETH moves on Binance, and bots trade Optimism's pools until the price there catches up. The swaps chase Binance. A minute after a swap, they match Binance's next move only 48% of the time, a coin flip. That chasing is what keeps Optimism's pools priced right. Contracts that follow Binance at least 70% of the time carried three-quarters of that June flow.

The flow is concentrated. Three contracts, each used by 20 wallets or fewer, carried a quarter of the year's trading on their own.

07 · What the count missesOne pool hides its trades

Bots also run the one pool that DEX trackers cannot see. Uniswap V4 lets a pool attach a hook, a small contract that runs its own code on every swap. Most hooks change a fee or add a check. One ETH/USDC pool on Optimism, opened in late July, has a hook that does the trading itself. Trackers read the swap events a pool writes, and this pool writes zero for every amount. To every tracker, ours included, its trades look empty.

The USDC still moves, and the chain records that. Following it gives about $268 million of trades between late July and early September. Counted in, it would add about 16% to the volume from July to September. One bot contract sent four in five of those swaps.

Five other V4 pools are inside our total but look wrong. Each paired ETH with a new token at a fee near zero and lived for 2 to 6 days. Together they did $38.5 million. Almost all of it came from wallets that only bought or only sold, and the ETH that flowed in matched the ETH that flowed out. That is the pattern of wash trading. We kept them in, because our total counts every swap, and they move it by under 1%.

08 · So, is anyone trading on Optimism?Yes. Mostly not people.

Optimism still clears about $20 million of trades per day. It held up better than Base or Arbitrum this year. It trades mainly ETH and dollars, and bots keep those prices in line with Binance.

What it lost is the crowd. Early in the year, most of that crowd came through bridge apps and traded pocket change, while smart accounts churned stablecoins at almost no fee. Both faded after March, and nothing took their place. Since July, about 20,000 to 23,000 wallets per month trade on Optimism, and people put under $3 million per day through its pools.

For a trader who wants to swap ETH and stablecoins cheaply, at prices kept close to Binance, that still works. For anyone looking for a busy market full of new tokens and new people, Optimism in 2026 is not it.

09 · How we measuredWhat counted, what did not, and why

We rebuilt every Optimism swap from Bitquery's raw on-chain data and checked it against a public Optimism node. Each swap is priced from the side we can price. Stablecoins count as $1. ETH, bitcoin and OP use Binance's hourly market prices.

What could go wrongHow we handled it
Failed tradesOnly swaps that went through count. Failed transactions can still leave swap records behind. On one day we checked, they made up a fifth of them.
Fake poolsA pool counts only if the DEX it claims to belong to really created it. Copycat pools that claimed quadrillions of dollars were dropped.
Tokens with no price0.5% of swaps trade only tokens we could not price. They are left out. At their own market prices they come to about $3 million.
Bots against peopleA contract built for bots is one that is not a known router, wallet, bridge or vault and makes more than 100 trades per day or arbitrages in a fifth of them. The table below shows how the bot share moves when these limits change.
Smart accountsSeveral smart accounts can share one transaction. Each swap is credited to the account that made it.
Base and ArbitrumMeasured from Bitquery's DEX data with the same prices. NFT sales are removed, and so is every pool that traded a token used to print fake trades. Pools that only look like wash trading stay in, as they do on Optimism. On Optimism, this method comes in 4% to 11% below our full rebuild.
The hidden poolEstimated from its USDC flows and kept out of every total.
Bot contract20 per day100500
Any trade85.5%77.5%67.7%
Most dollars83.2%75.2%67.7%
Bots' share of the year's volume under each version of the rules. Columns: how many trades on an active day make a wallet a bot; 100 is our default. Rows: one trade through a bot contract makes a wallet a bot, or only when most of its dollars go through one.

10 · The recordThe addresses behind the story

RoleAddress
Largest pool, USDC/WETH0x4789…ba25
OKX DEX router0x6733…0d19
Main paymaster0x7777…834c
ERC-4337 EntryPoint0x0000…a032
Relay approval contract0xccc8…15be
Uniswap V4 PoolManager0x9a13…3ec3
Hook of the hidden pool0x8863…8088
Bot contract, 20 callers0x1ef3…39e9
Bot contract, 3 callers0x5e6a…37b9
Bot contract, 3 callers0x9837…4c69
Top arbitrage contract0x4661…c6b8
Run it yourself

Ask these questions in plain English

Every count above came from data anyone can query. The Bitquery MCP server puts it behind an AI assistant, so you can ask how much a chain's DEXs traded this week, which pools carried it, or which wallets trade like bots, without writing the query yourself.

Rank Optimism DEX pools by volume for any weekCompare DEX volume across Optimism, Base and ArbitrumFind the wallets behind a pool's tradesFollow a smart account's swaps through its bundle
Explore Bitquery MCP Figures measured September 21, 2026, against Bitquery's Optimism, Base and Arbitrum data. Written by Bitquery Research; AI tools ran the queries and drafted the text, and every figure was worked out again from the raw data and checked before we published.
Scope, limits and attribution

This article is provided for informational and educational purposes only and reflects analysis of publicly available on-chain data as of the dates indicated. It does not constitute legal, financial, compliance, or investment advice, and nothing in it is a recommendation to buy, sell, or hold any token or asset.

The findings describe decentralized exchange swaps recorded on Optimism between January 1 and September 20, 2026, and, for comparison, swaps recorded in Bitquery's DEX-trades data for Base and Arbitrum over the same months. The split between bots and people is a rules-based classification of wallet and contract behaviour; it is an inference, it shifts with the thresholds shown, and a wallet labelled a bot may be run by a person. Dollar values depend on the prices used and exclude swaps we could not price. The estimate for the pool whose swap events log zero amounts is drawn from its token flows and may be incomplete.

References to OKX, Relay, Velodrome, Uniswap, the paymaster contract and any other named party describe on-chain routing and contract calls, not the intent or knowledge of any company or person. Terms such as volume farming and wash trading describe trading patterns visible on the chain; they are not allegations that any named party acted unlawfully or improperly.

Nothing herein should be relied upon as a definitive determination of fact. Readers should conduct their own independent verification before taking any action. The authors and publisher accept no liability for any loss or damage arising from reliance on this material. All trademarks and company names are the property of their respective owners.