solana sandwich attacks Hit Memecoins for $29.7M, Bitquery Says

solana sandwich attacks are eating memecoin buyers alive, and Bitquery just put a hard number on the bill. Across 11 months of Solana swaps, sandwich bots ran 6,850,144 attacks and took $29.7 million before costs. Almost all of that came out of Pump.fun memecoin trades, and wide slippage settings made the hit rate much worse.

What actually happened

The Bitquery sandwich attack investigation, with on-chain figures verified October 11, 2026, walked 11 months of Solana DEX history to measure how often bots buy just before a victim and sell just after. The headline take is $29.7 million, measured as 309,010 SOL priced at the SOL minute of each attack.

The mechanics are simple. A bot buys a coin just before your trade, which pushes the price up. Your trade fills at the worse price, still inside your slippage limit. The bot sells right after and keeps the difference. In one September 16 example Bitquery traced, a wallet spent about $240 in SOL on a memecoin and got 6% fewer coins because of a bot buy a fraction of a second earlier. The bot made about $11 and kept about $8 after fees.

Single attacks are small. The median sandwich made the bot $2.33, and one in ten made more than $12. In a checked sample of 63 attacks, the bots kept 68% of the gross after fees and tips, and 18 of the 63 lost money once costs were paid.

Where the money comes from

PumpSwap took 62% of the sandwiches and the Pump.fun launch curve 34%. Meteora had 3.1% and Raydium 1.0%. Memecoins suit the bots: small pools move a lot on one buy, and buyers tend to allow wide slippage because prices jump around.

The pace exploded. Bitquery's monthly averages went from about 3,400 attacks a day in November and December 2025 to 51,900 a day in September 2026, and the first week of October was busier still. Pump.fun swaps per day only tripled, so the attack rate per swap rose about six times.

Plenty of victims were not humans. About 100,000 victim wallets were bots by Bitquery's counts, such as snipers and very high-volume wallets, and 39% of the take came from bot trades. The other 61%, $18.1 million, came from everyone else.

Bitquery's count is a floor. It only counts wallets where sandwiches make up at least a quarter of trading. Adding busy wallets that also sandwich would roughly double both the attack count and the take, to between $58.6 million and $63.5 million.

Your 20% slippage setting is not a safety net. It is a tip jar for bots.

Why solana sandwich attacks keep winning

Solana has no public mempool like Ethereum, so bots learn about trades in other ways. A September 23 study, "No Place to Hide," by researchers from ETH Zurich, the University of Lisbon, Flashbots and Category Labs, found 28.0 million sandwich attacks on Solana against transactions that were meant to be protected from front-running, over three years. Most Solana sandwiches no longer sit tightly around the victim; the bot's buy and sell can land a few blocks apart.

Bot wallets also churn fast. Half of Bitquery's sandwich wallets were active for seven days or fewer, which makes blocking them one address at a time pointless. Faster slot times help a bit, as our Solana 200ms slots piece explains, but speed alone does not fix ordering. Related coverage of Jupiter Ultra and sandwich risk is also in CryptoBriefing's study write-up.

What to do as a reader

Tighten your slippage. On Pump.fun coins, Bitquery found swaps allowing 1% to 3% were caught 8 times per 10,000, against 148 per 10,000 at 20% to 50%. Tight settings are not magic, though: swaps under 0.5% were hit more often than 1% to 3%, mostly in a March to April burst, and very tight limits mean more failed trades.

Use routes with private submission when you can. Jupiter's own risk page says Ultra Mode includes MEV protection through private transaction submission, which reduces the risk but does not eliminate it, and Manual Mode has none. If you trade new launches, also read our Pump.fun custom pairs explainer.

solana sandwich attacks profit from impatience. Smaller orders, tighter limits and fewer panic buys at launch all shrink the gap they live in. Nothing here is investment advice.