PropAMMs and the Decentralization Question
Published 6/14/2026, 3:09:38 PM
The research reveals a meaningful tension between propAMM architecture and DeFi's decentralization principles, though the evidence is primarily drawn from institutional research without publicly citable URLs.
PropAMM Architecture
PropAMMs operate as a two-part system: an offchain pricing engine that monitors CEX order books (Binance, Bybit, Coinbase) and computes fair value, combined with an onchain execution program that maintains proprietary liquidity. The architecture enables ultra-low latency oracle updates — typically every 100 milliseconds on Solana — with compact binary payloads used to execute trades with asymmetric pricing.
Centralization Risks
| Decentralization Concern | Severity | Evidence |
|---|---|---|
| CEX Price Feed Reliance | High | Core pricing derived entirely from centralized exchanges |
| Closed-Source Implementations | High | All propAMMs on Solana are closed-source; users cannot verify pricing logic |
| Operator-Only Liquidity | High | 100% of liquidity from professional MMs; no permissionless LP deposits |
| Aggregator Concentration | High | Jupiter routes 80-95%+ of Solana aggregator volume |
Scale of concentration: PropAMMs now facilitate over 60% of Solana's onchain spot trading volume, with HumidFi alone capturing 47.1% of propAMM volume as of mid-June. Total daily volume consistently exceeds $1 billion with peaks of $1.93 billion.
Documented Execution Issues
Research from 0x on Base (where 5 propAMMs drive 40-50% of Swap API volume) reveals systematic worse-execution patterns:
- Flashblock manipulation: Operators publish competitive prices in the final ~200ms of a block to influence aggregator routing snapshots, then reprice worse when the trader's transaction settles. This pattern occurs 97.4% of the time in the final 10% of blocks.
- Random spread fluctuations: Tight spreads (~2 bps) snap wide to 8-16 bps between quote and settlement.
- Phantom liquidity: LP deposits millions in the last flashblock, collects full block rewards, then withdraws in the first flashblock of the next block.
These patterns extracted an estimated ~$500,000/month from traders at $1B monthly volume with consistent 5 basis point overquoting.
The Oracle Spectrum
| Oracle Type | Decentralization Level | Manipulation Resistance |
|---|---|---|
| Built-in TWAP (Uniswap v4) | High | High for liquid pools, vulnerable to multi-block attacks in PoS |
| Truncated Oracle | High | Caps tick movement at 9,116/block; requires sustained multi-block attack |
| Chainlink Network | Medium-High | Decentralized node operators, but data sourced from institutional providers |
| CEX Aggregated Feeds | Low | Centralized infrastructure, front-running possible |
Academic research from the University of Basel analyzing 150 million+ observations across 40 Chainlink Price Feeds over 18 months found economically significant deviations systematically related to oracle accuracy configurations and market stress conditions. Chainlink secures over 70% of the DeFi ecosystem for lending markets, stablecoins, and derivatives.
Assessment
Yes, CEX reliance creates meaningful centralization risks, but the trade-off is not binary:
-
PropAMMs create architectural centralization through CEX price feed dependency, closed-source implementations, and operator-only liquidity — directly contradicting DeFi's principles of transparency, permissionlessness, and verifiability.
-
The systemic concentration is significant — with a single blockchain (Solana), single aggregator (Jupiter at ~93.6% market share), and centralized oracle feeds handling the majority of onchain spot volume, cascading failures are a credible risk.
-
The execution quality vs. decentralization trade-off is real — protocols can choose positions on the decentralization spectrum based on use case requirements, but CEX-competitive execution comes at the cost of architectural centralization.
What Remains Open
The critical question the ecosystem faces: Can propAMMs evolve toward greater decentralization without sacrificing the performance advantages that make them viable? Without structural changes — open-sourced implementations, permissionless liquidity provision, or decentralized oracle alternatives — propAMMs risk transforming DeFi into a more efficient version of the same centralized market structure they're meant to disrupt.
Note on citations: The claims above are supported by institutional research from 0x Research, Jump Crypto, and the University of Basel, but the skill outputs did not contain publicly citable https:// URLs. The evidence should be verified against primary sources before use in formal contexts.
Claim Resolution Status:
- c1 (propAMMs rely on CEX feeds): UNRESOLVED — no citable URLs in outputs
- c2 (CEX dependency creates centralization risks): UNRESOLVED — narrative evidence from institutional research without linked sources
- c3 (trade-offs well-documented): UNRESOLVED — documentation is narrative-based without explicit URLs
Suggested Next Steps
-
Verify primary sources — Locate and cite the original 0x Research and Jump Crypto reports on propAMM architecture to obtain citable URLs for the claims above.
-
Monitor execution quality — Set up a recurring analysis to track whether the flashblock manipulation patterns documented on Base are emerging on Solana propAMMs, given the similar architecture.