Solana preprint finds wider spreads outside aggregators at five of six propAMMs
In brief
- Solmaz, Heimbach and Milionis preprint, posted Sept. 29, studied proprietary AMMs on Solana, Base and Monad.
- Five venues showed wider non-aggregator spreads in 68% to 83% of matched cells.
- ZeroFi was the exception, posting a small negative matched median premium.
- Figures measure realized round-trip price gaps, not individual swap fees.
- Data spans Sept. 1, 2025, to Aug. 31, 2026; findings reflect a historical association.
What the preprint measured
The paper looks at proprietary automated market makers (propAMMs) on Solana, Base and Monad. A propAMM is a venue where one operator provides inventory and can update prices directly, without a trade having to move the pool's reserves. That's the key design choice. Operators can push price updates without executing a swap, which reduces their exposure to stale-price losses.
On Solana, the authors compared aggregator and non-aggregator flow within the six largest measured propAMMs for SOL/USDC. (Jupiter is one such routing provider on Solana.) Observations run from Sept. 1, 2025, through Aug. 31, 2026, with later start dates for newer deployments.
To estimate spreads, they paired consecutive opposite-side fills in the same pool, within 25 slots and with trade sizes within a factor of two. There couldn't be an intervening operator quote update, and both legs had to belong to the same routing class. The researchers then compared classes within cells defined by pool, day and trade-size bucket.
Five venues, one exception
At the first five venues, non-aggregator spreads were wider in 68% to 83% of matched cells, with reported exact sign-test p-values below 0.001.
ZeroFi went the other way.
It showed a small negative matched median premium, so the pattern didn't hold at every venue. Tessera's overall median spreads were 0.9 and 0.93 basis points, and its median matched premium was 0.1 basis points. In the Tessera implementation on Base, the researchers identified six fee components: base, age, priority, mode, default and address.
What it doesn't show
The limits matter here. These are round-trip price gaps, not individual swap fees. The paired fills aren't necessarily one customer's actual buy-and-sell sequence, and the non-aggregator label doesn't identify ordinary retail customers. Matching also doesn't turn the comparison into a randomized test of what happens when the same person changes interfaces.
For defi traders on Solana, that's the caveat to keep in mind: the preprint describes a past association between routing class and realized spreads, not a measurement of what any app delivers on a swap today.
Frequently asked questions
What is a propAMM?
A proprietary automated market maker (propAMM) is a liquidity venue where one operator provides inventory and can update prices directly, without waiting for a trade to change the pool's reserves. Operators can send price updates without executing a swap, which reduces their exposure to stale-price losses.
Does the study show non-aggregator traders paid higher swap fees?
No. The measured figures are realized round-trip price gaps, not individual swap fees. The paired fills need not be one customer's buy-and-sell sequence, and the non-aggregator class doesn't identify ordinary retail customers.
Did the wider-spread pattern hold at every Solana venue studied?
No. At the first five venues, non-aggregator spreads were wider in 68% to 83% of matched cells, with reported p-values below 0.001. ZeroFi was the exception, showing a small negative matched median premium.
