A pre‑print study released on September 29, 2026 examined proprietary automated market makers (propAMMs) on Solana, Base and Monad, focusing on SOL/USDC swaps between September 1, 2025 and August 31, 2026. The authors – Ozan Solmaz, Lioba Heimbach and Jason Milionis – found that trades routed outside aggregators incurred noticeably wider realized round‑trip spreads at five of the six largest Solana propAMMs, suggesting that simply accessing a pool does not guarantee comparable trading terms.

SOL / USD
Source: CoinGecko. Prices are indicative.

Methodology and Data Set

The researchers paired consecutive fills on opposite sides of the same pool within 25 slots, limiting trade sizes to a factor of two and ensuring no intervening quote updates. Both legs of each pair had to belong to the same routing class—either aggregator‑derived or non‑aggregator flow—allowing a cell‑by‑cell comparison of spreads by pool, day and size bucket. This approach isolates the impact of routing while controlling for pool‑specific and temporal factors, though it does not track a single user’s complete buy‑and‑sell journey.

Key Findings on Spread Differences

Across the five venues where non‑aggregator spreads were wider, the matched‑cell analysis showed that 68 % to 83 % of cells exhibited larger spreads for non‑aggregated trades, with sign‑test p‑values below 0.001, indicating strong statistical significance. Median spreads for the propAMM Tessera were 0.9 basis points (bp) for aggregator flow versus 0.93 bp for non‑aggregator flow, with a median matched premium of 0.1 bp favoring aggregators. ZeroFi was the sole outlier, showing a slight negative premium for non‑aggregator flow. The study emphasizes that these are round‑trip price gaps—not swap fees—and that the routing class itself influences pricing.

Why Aggregators May Offer Better Terms

Aggregators such as Jupiter scan multiple liquidity venues and can split orders across them, selecting the most favorable quotes. PropAMM operators, by contrast, set prices directly and can update quotes without executing a trade, reducing exposure to stale prices. The study suggests that aggregator routing mitigates stale‑price risk by quickly accessing fresher quotes, while non‑aggregated routes may encounter delayed updates, leading to wider spreads. On Base, the authors identified fee components (base, age, priority, mode, default, address) that can vary based on the transaction sender, further illustrating how proprietary venues embed caller identity into pricing.

What to Watch Next

Future research should aim to capture quote‑to‑fill dynamics on Solana, as the current study lacked historical quote data for that chain. Monitoring whether newer propAMM deployments adopt similar fee structures or improve price‑update latency will be crucial. Additionally, wallet‑level fees—such as Phantom’s 0.85 % swap charge—remain separate from on‑chain spread measurements and could affect net execution costs for retail users. Observers should also track any changes in Jupiter’s fee schedule or routing algorithms, as these could shift the comparative advantage highlighted by the study.