Author: Kent Lin
A hacker sitting on stolen assets worth billions burned roughly $119 million just by swapping tokens. The cash didn’t vanish. It was redistributed in plain sight to the fastest hands on Ethereum.
This ChainCatcher piece traces that path through arbitrage bots, block builders, and validators. On Oct. 30, one Ethereum block builder pulled in more fees in a single day than Tether, based on the figures cited here. DefiLlama first tagged the number as dirty data and blocked it.
$119 million lost in rushed on-chain swaps
The article’s read is simple: the hacker chose speed over price. When billions in stolen funds might get frozen, time matters more than execution quality. In that kind of rush, price discipline goes out the window.
So the hacker made the kind of trade ordinary traders try very hard not to make: huge size, on-chain, urgent, and at whatever price the pool would give. Bitquery put the damage at about $119 million versus market prices.
The ugliest point hit at 21:09 UTC, which was 5:09 a.m. in Singapore and Beijing on Oct. 30. In Uniswap’s main WETH/USDT pool, ETH was bought at a 38% premium to the market price. The buyer willing to pay that was the hacker.
Selling was painful too. Tokens such as UNI, PEPE, and ENA were dumped at discounted prices, giving up another $24.8 million.
Set against Binance’s estimate of a $6.3 billion impact, the loss was under 2%. Still, nine figures moved from a desperate seller to whoever reacted fastest, and the whole thing played out on-chain for everyone to watch.
How the money moved down the stack
The article splits the roughly $119 million across a few layers.
- Arbitrage traders and bots took about $60 million after paying gas and tips to builders.
- Because several bots were chasing the same dislocated pools, they had to pay block builders to get to the front of the line. That sent about $30 million to builders and validators.
- Builders then passed most of that along to validators, which the article says matches Bitquery’s wording.
What the builder kept was the actual business income.
Titan’s one-day revenue dwarfed its normal pace
The article says Titan posted $13.13 million in revenue over a 24-hour stretch after validator payouts. For context, Titan made $4.66 million in all of September, or about $155,000 a day on average. So one day came out to about 2.8 times its entire September revenue and roughly 84 times its average daily pace that month.
The piece spends a while on method because the numbers do not match, and they are not meant to. Bitquery tracked only the hacker’s own fund flows. DefiLlama counted everything Titan processed in that same window. 0xngmi also said panic across the wider market created arbitrage chances beyond the hacker’s swaps.
That’s why the totals are different. Nick Research put Titan’s fees at $14.91 million, while DefiLlama’s 24-hour leaderboard showed $24.68 million.
The article also fixes a common mix-up: $24.68 million is fees, not revenue. Revenue is what remains after subtracting validator payouts.
Fee rankings showed how much value reached the ordering layer
Using the DefiLlama 24-hour fee table cited in the article, the top entries were:
- Titan Builder: $24.68 million
- Tether: $16.99 million
- Flashbots: $13.73 million
- Uniswap: $12.89 million
- Ethereum: $10.82 million
- Hyperliquid: $6.36 million
The article lingers on the fifth line. A company that packages Ethereum blocks took in more fees than Ethereum itself.
But the comparison is not perfectly one-to-one. Builder fees include MEV payments, while Ethereum’s figure reflects gas. Even so, the ranking makes the direction of value movement pretty hard to ignore.
Across seven days, the pattern was still there. Titan recorded $25.95 million in revenue, trailing only Tether at $103 million and Circle at $40.88 million. Hyperliquid logged $14.10 million over that same stretch.
The broader point is blunt. Crypto attention usually sits at the application layer, especially exchanges and perpetual DEXs. But when a $6.3 billion hole opens, a big chunk of the money can end up with the layer that decides transaction order.
Speed gaps between builders widened sharply under stress
The article argues that builders are easy to overlook in calmer conditions because their earnings rarely attract much notice. In a crisis, that changes fast. Every bot is trying to hit the same broken pool first. Whoever sees more order flow and pushes blocks out faster wins the auction.
On that same leaderboard, BeaverBuild posted $2.56 million in fees, while Titan posted $24.68 million. Same chain. Same event. Almost a tenfold gap.
The article adds that among the three builders on the board, about 60% of combined fees went to Titan alone, on the very day block-ordering value climbed into the tens of millions of dollars. Under stress, concentration can get extreme.
The author connects this to work at Optimum (@get_optimum), which is focused on how fast block data moves across Ethereum. The article says Optimum CEO Muriel Médard is an MIT professor and a co-inventor of RLNC, or random linear network coding, and that the company uses that coding method to speed block propagation between validators. It says the system went live on Ethereum mainnet on Oct. 20.
The author says Optimum is not a builder and does not handle ordering. Even so, the episode made one thing concrete: once ordering has real value, a delay of a few hundred milliseconds can convert straight into revenue, and here the scale hit tens of millions of dollars in a single day.
Flashbots booked fees, but no revenue
During that same 24-hour period, Flashbots showed $13.73 million in fees, while DefiLlama listed its revenue at $0.
By DefiLlama’s definition, that ETH was paid by users to block-producing validators and merely passed through Flashbots on the way. Titan, by contrast, kept $13.13 million in revenue.
The article puts it like this: Flashbots looked more like road infrastructure, sending value back to the validators that maintain the chain. Titan looked like a business with a margin. And it kept that margin.
The author does not say either model is wrong. The point is narrower than that. Ethereum has spent years trying to push MEV value back to validators, yet on this day the most profitable actor in the stack was the one that kept more of the money for itself.
Binance absorbs the loss as debate shifts to builder policy
The article says Binance has already said that SAFU and the company’s own funds will fully cover user losses. In that framing, users are not paying the “panic tax.” Binance’s balance sheet is.
The author raises another angle he says has received little attention: every dollar the hacker lost to slippage was a dollar that did not make it into a mixer. In that sense, the arbitrage bots did something useful that night, even if they did not mean to.
The article quotes Qin Zong, @qinbafrank, asking: “Does this count as helping launder the hacker’s dirty money?” The author says that question stopped him because both sides have an argument. The builders earned record revenue that day, and those same builders were also the ones packaging the hacker’s swaps into blocks.
A bet on written policies before the end of 2026
The article ends with a prediction. If another exchange gets drained on this scale, major builders will probably face public pressure to reject the attacker’s transactions during the first few hours.
Some may comply. Others may take the order flow and the fees that come with it.
The author says he would bet that before the end of 2026, at least one major builder will publish a written policy on handling stolen funds. He also expects Ethereum social media to break into opposing camps when that happens.
From the point of view of network-layer operators, the article adds one more judgment: the few hundred milliseconds many people used to shrug off will probably get more expensive over time.


