Jupiter Ultra Shows Lower Sandwich Attack Exposure Than Other Apps

Research from Category Labs, ETH Zurich, Flashbots, and Lisbon University revealed that Jupiter (JUP) Ultra users experienced lower exposure to sandwich attacks compared to other trading applications on Solana. The study analyzed protected order flows across Ethereum (ETH), Solana (SOL), Tron, Base, Arbitrum, and Monad from July 1, 2023, to June 30, 2026, with results published September 23.
The researchers noted that even within protected order flows, transaction information could still be exposed. Sandwich attacks occur when bots detect pending user transactions, buy before the user's trade executes, and sell immediately after, profiting from price differences while users face unfavorable rates.
On Solana, 28 million sandwich attacks were identified, involving 8,631 bots generating $383.4 million in gross revenue (approximately ₩519.9 billion), with net profits of $345.2 million ($468.1 billion) after fees. These figures are limited to the researchers' detection criteria and protected order flow analysis scope, not representing total Solana transaction volume.
The study measured 'excess ratio,' indicating how frequently users on a specific app were targeted compared to a baseline. Jupiter Ultra's overall excess ratio was 0.7, below the baseline, while a single victim-based calculation showed 2.0—still lower than competitors like Axiom (18.9) and Photon (11.1). BullX and GMGN also exceeded 10.
The difference highlights how order routing and timing of transaction exposure affect attack frequency across the same blockchain. Jupiter explained that Ultra V3 offers Iris Meta Aggregation, JupiterZ, dynamic slippage estimation, predictive execution, and its own transaction transmission infrastructure to improve security. The company claimed a 34x improvement in sandwich attack defense over competing trading terminals as of October 2025.
However, the researchers emphasized that protected order flows do not fully prevent sandwich attacks. Solana showed exposure at both validator and application layers, with bots splitting transactions across multiple accounts or executing across block boundaries. While Jupiter Ultra users had relatively lower attack exposure, this does not mean sandwich attacks have been eliminated. Factors including order routing, timing of transaction disclosure, slippage settings, and validator/application structure all influence exposure levels.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: 주피터 울트라, 샌드위치 공격 초과 비율 0.7