Are propAMMs sacrificing decentralization by
Published 6/14/2026, 3:12:07 PM
Answer
Based on the research, PropAMMs are indeed sacrificing decentralization in price discovery by relying on CEX price feeds — but this is a deliberate trade-off rather than an unintended compromise.
Claim-by-Claim Assessment
c1: PropAMMs use CEX price feeds as primary price discovery mechanism
Status: Supported, with URL citations
The architecture confirms CEX dependency. PropAMMs operate a two-part system: an off-chain pricing engine continuously monitoring CEX order books (Binance, OKX, Coinbase spot markets) and an on-chain execution program that receives price updates [Source: https://jumpcrypto.com/research/propAMMs-and-the-next-chapter-of-permissionless-market-structure].
This design enables spreads as low as 0–0.5 bps on SOL/USDC, rivaling CEX execution quality [Source: https://www.gate.io/research/analysis/2388]. However, this tight pricing comes from anchoring to external CEX markets rather than on-chain discovery.
c2: CEX reliance introduces centralization risks
Status: Supported, with specific mechanisms documented
Four sources confirm centralization vectors:
| Risk | Mechanism | Evidence |
|---|---|---|
| Oracle manipulation | Aggregator spoofing — publish attractive price in last Flashblock (~200ms), reprice worse in first Flashblock of next block | 97.4% of final price updates land in last 10% of block [Source: https://0x.org/research/propAMM-Shenanigans] |
| Spread widening | Quote tight ~2 bps to win routing, then widen to 8–16 bps before settlement | ~$500,000/month extracted from traders at $1B monthly volume [Source: https://0x.org/research/propAMM-Shenanigans] |
| Permissioned integration | Each PropAMM must individually approach aggregators (Jupiter holds ~80–95%+ of Solana aggregator market share) | Adding to router is a closed process [Source: https://jumpcrypto.com/research/propAMMs-and-the-next-chapter-of-permissionless-market-structure] |
| Single-source liquidity | 100% of liquidity comes from operator capital; no permissionless LP deposits | Closed liquidity model [Source: https://solana.com/research/understanding-proprietary-AMMs] |
Gap: Direct censorship incidents or oracle manipulation events causing sustained outage not documented in these sources.
c3: Trade-off between price accuracy/speed (CEX) and decentralization (DEX/TWAP)
Status: Confirmed, with quantified performance gap
| Dimension | CEX-Feed PropAMMs | Traditional AMMs |
|---|---|---|
| Price discovery | Off-chain via CEX feeds | On-chain via arbitrage |
| Execution speed | ~100ms per slot | Block-contingent |
| Spread | 0–0.5 bps (SOL/USDC) | Typically 5–50+ bps |
| Decentralization | Partial (execution only) | Full (discovery + execution) |
Execution quality on Solana PropAMMs matches or beats CEXs: 0.33–1.36 bps from best CEX mid (March 2026 data) [Source: https://solana.com/research/understanding-proprietary-AMMs]. Traditional AMMs cannot achieve this because on-chain price discovery is inherently slower than CEX monitoring.
Gap: No quantitative data on what percentage of PropAMMs use CEX feeds versus alternatives (TWAP, DEX oracles). The trade-off is qualitative rather than quantified in available research.
c4: Some PropAMMs use hybrid or decentralized approaches
Status: Partial — limited documented examples
Humidify represents the exception, pursuing what they call "abstinence trading" — building price models using only on-chain data to eliminate CEX dependencies. Their ORE token test case aims to become a price discovery venue using purely decentralized trading activity [Source: https://jumpcrypto.com/research/propAMMs-and-the-next-chapter-of-permissionless-market-structure].
Gap: No documented PropAMMs explicitly combining CEX feeds with decentralized oracle networks in a hybrid approach. More implementation-specific examples are needed.
The Nuance
The "decentralization sacrifice" framing is somewhat misleading. PropAMMs achieve:
- Decentralized execution: Settlement happens on public ledgers with immutable, independently verifiable records
- Centralized price discovery: Anchoring to CEX markets rather than on-chain discovery
Every oracle update and trade is recorded immutably on-chain, and execution quality is independently computable against NBBO — attributes that exceed current Reg NMS reporting requirements [Source: https://solana.com/research/understanding-proprietary-AMMs]. The trade-off is not between "decentralized" and "centralized" but between trust-minimized execution (achieved) and fully trustless price discovery (not attempted due to oracle problem constraints).
Conclusion
Yes, PropAMMs sacrifice decentralization in price discovery by relying on CEX feeds — but this is a conscious engineering decision that delivers CEX-comparable execution quality on-chain. The real trade-off is execution decentralization versus price discovery decentralization; execution is decentralized, price discovery is not. What remains unresolved is whether the documented manipulation patterns (spoofing, spread widening) represent acceptable costs for the execution quality gains, and whether hybrid approaches like Humidify's can achieve comparable pricing without CEX dependencies.
Suggested Next Steps
-
Deep-dive on manipulation patterns: Pull the 0x "PropAMM Shenanigans" report in full to assess whether the $500k/month extraction rate is structurally embedded or fixable through protocol changes.
-
Compare Humidify's performance: If ORE/USDC or similar pairs are live, compare Humidify's execution quality (spread, slippage) against CEX-feed PropAMMs to test whether CEX-free pricing is viable.