AMMs aggregate via liquidity pools. They replace order books with a deterministic pricing function, like x*y=k, which creates a continuous price curve from pooled capital. This is a primitive form of information synthesis, where the price signal emerges from the ratio of two token reserves.
prediction-markets-and-information-theory
Blog
Why AMMs Are a Primitive Form of Information Aggregation
Constant function market makers like Uniswap aggregate liquidity preferences into a price, but lack the expressiveness to aggregate beliefs about future events. This post dissects the information theory gap between AMMs and true prediction markets.
introduction
THE PRIMITIVE
Introduction
Automated Market Makers are a rudimentary, on-chain mechanism for aggregating liquidity and price information.
This model is informationally inefficient. The price only updates on trade execution, making it a lagging indicator. Unlike UniswapX or CowSwap which aggregate intents off-chain, a standard AMM like Uniswap V3 cannot see latent demand before a trade settles.
The evidence is in MEV extraction. Over $1.2B in MEV has been extracted from DEXs, primarily from arbitrage bots correcting AMM prices after external market moves. This arbitrage is the cost of the AMM's slow information loop.
thesis-statement
THE DATA
The Core Argument: Liquidity ≠Belief
Automated Market Makers aggregate capital, not conviction, creating a fundamental mispricing of information.
AMMs aggregate capital, not information. Their pricing logic is a deterministic function of pool reserves, reacting to trades after they occur. This creates a lag where price discovery is a byproduct of liquidity depletion, not a leading signal.
The oracle problem is inverted. Protocols like Chainlink and Pyth import external data on-chain. An AMM's price is an output of internal mechanics, making it a reactive price reporter, not a proactive price discoverer*.
Liquidity providers are passive signalers. Depositing into a Uniswap V3 pool expresses a belief in a price range, not a specific value. This capital efficiency divorces liquidity commitment from precise price conviction.
Evidence: Over 90% of DEX volume occurs on AMMs, yet perpetual futures on dYdX and GMX consistently lead spot price movements. The market for beliefs is structurally separate from the market for swaps.
key-trends
FROM PRIMITIVE TO SOPHISTICATED
The Information Aggregation Spectrum
Automated Market Makers are the first, most primitive layer of on-chain information aggregation, relying on static formulas rather than dynamic signals.
01
The Problem: Static Formulas, Blind to Reality
AMMs like Uniswap V2 aggregate liquidity but not information. Their price is a function of a constant product formula, not real-time market sentiment or external data feeds.
- Price Divergence: AMMs are perpetually behind centralized exchanges, creating arbitrage opportunities that extract ~$1B+ annually from LPs.
- Inefficient Capital: The 50/50 pool ratio forces overexposure to illiquid assets, locking up capital that could be deployed elsewhere.
$1B+
Annual Arb Loss
50%
Forced Allocation
02
The Solution: Concentrated Liquidity (Uniswap V3)
Uniswap V3 introduced programmable liquidity ranges, allowing LPs to express a price opinion. This is a step toward information aggregation by letting capital signal where it believes the market will trade.
- Capital Efficiency: LPs can achieve up to 4000x higher capital efficiency within a tight range.
- Active Management Burden: The 'solution' shifts complexity to LPs, requiring constant monitoring and rebalancing, turning them into quasi-market makers.
4000x
Capital Efficiency
High
LP Overhead
03
The Next Frontier: Oracle-Integrated AMMs (Curve, Maverick)
Protocols are integrating external price feeds to make the AMM formula itself dynamic. This moves aggregation from just liquidity to incorporating external market data.
- Curve's EMA Oracle: Uses a time-weighted average price to reduce slippage and resist manipulation for stablecoin pairs.
- Maverick's AMM: Allows LPs to auto-shift liquidity based on an oracle price, dynamically aggregating both capital and market price signals.
TWAP
Price Feed
Dynamic
Liquidity
04
The Limit: AMMs Cannot Aggregate Intent
Even advanced AMMs fail at aggregating user intent—the willingness to trade at a specific price across chains or time. This is the domain of intent-based protocols like UniswapX, CowSwap, and Across.
- Solving MEV: These systems batch and settle orders off-chain, finding optimal paths and extracting value for users, not searchers.
- Cross-Chain Native: They aggregate liquidity and settlement options across domains (e.g., via LayerZero, Chainlink CCIP), which a single AMM pool cannot do.
Intent-Based
Paradigm
Cross-Chain
Aggregation
deep-dive
THE DATA PIPELINE
Information Theory 101: From Uniswap to Hanson
Automated Market Makers are primitive information engines, aggregating trader signals into a single price feed.
AMMs are information aggregators. Every swap is a data point. The constant product formula x*y=k is a simple algorithm that processes these signals to output a new price state. This is a decentralized, albeit slow, form of consensus on asset value.
Uniswap's price is a lagging indicator. It reflects past trades, not future intent. This creates predictable arbitrage latency that MEV bots exploit. This inefficiency is the information gap that intent-based architectures like UniswapX and CowSwap target.
Robin Hanson's prediction markets are the logical endpoint. They formalize the AMM's role as an information processor. Where Uniswap aggregates liquidity, a Hanson-style market maker aggregates beliefs, paying for accurate data to converge on truth.
Evidence: The 2022 UST depeg. DEX pools like Curve 3pool were the first on-chain signals of imbalance, but their reactive pricing lagged the off-chain collapse. A predictive, information-seeking AMM would have priced the risk earlier.
INFORMATION AGGREGATION PRIMITIVES
AMM vs. Prediction Market: A Feature Matrix
Comparing the core mechanisms by which Automated Market Makers and Prediction Markets aggregate and price information, revealing AMMs as a primitive, latency-sensitive subset.
| Feature / Metric | Automated Market Maker (e.g., Uniswap v3) | Central Limit Order Book (e.g., Binance Spot) | Prediction Market (e.g., Polymarket, Kalshi) |
|---|---|---|---|
Primary Information Source | Passive liquidity & arbitrage latency | Active trader orders & market makers | Crowd-sourced beliefs on future events |
Price Discovery Mechanism | Bonding curve (x*y=k) & external arbitrage | Direct order matching at limit prices | Market probability derived from token price (e.g., $0.75 = 75% chance) |
Information Latency Sensitivity | High (seconds-minutes for arb) | Very High (milliseconds) | Low (hours-days for event resolution) |
Expresses Uncertainty Quantitatively | False (only expresses current price) | False (only expresses current price) | True (price = probability) |
Native Time Dimension | False | False | True (expiry date) |
Settlement Type | Continuous (spot swap) | Continuous (spot trade) | Binary or Scalar (on event outcome) |
Typical Fee for Aggregation | 0.01% - 1% LP fee | 0.1% taker fee | 2% - 10% market creator fee |
Capital Efficiency for Info Role | Low (locked capital for all prices) | High (capital deployed at specific prices) | High (capital only at risk on specific outcome) |
counter-argument
THE PRIMITIVE AGGREGATOR
Steelman: Aren't AMMs Evolving?
Automated Market Makers are a primitive, lossy form of information aggregation that is being superseded by intent-based architectures.
AMMs are lossy aggregators. They compress all user preferences into a single, static price curve, discarding nuanced information like urgency, size, and cross-chain intent.
Intent protocols are the evolution. Systems like UniswapX and CowSwap separate expression from execution, allowing solvers to aggregate and optimize orders across venues, reducing MEV and improving prices.
The primitive is the liquidity pool. An AMM's constant function formula is a blunt instrument compared to the dynamic, multi-venue optimization performed by solver networks in intent-based systems.
Evidence: UniswapX now routes over 50% of Uniswap's swap volume, demonstrating market preference for aggregated liquidity and MEV protection over direct AMM interaction.
protocol-spotlight
FROM AMMS TO INTENTS
Beyond the Primitive: Next-Gen Aggregators
Automated Market Makers are a primitive form of information aggregation, limited to on-chain liquidity and price. Next-gen aggregators solve for the user's intent, not just the trade.
01
The Problem: AMMs Aggregate Only Price
AMMs like Uniswap V3 are single-dimensional, exposing users to MEV and failing to access the best execution across venues.\n- Limited to on-chain liquidity in its pool\n- Vulnerable to sandwich attacks and frontrunning\n- Ignores off-chain liquidity from CEXs or private market makers
~$100M+
Annual MEV Loss
1
Dimension (Price)
02
The Solution: Intent-Based Architectures
Protocols like UniswapX, CowSwap, and Across shift the paradigm. Users declare a desired outcome (intent), and a network of solvers competes to fulfill it optimally.\n- Access to all liquidity: on-chain, off-chain, and private\n- MEV protection: Solvers internalize value, turning extractive MEV into better prices\n- Gasless experience: Users sign messages, solvers handle execution and gas
~$10B+
Volume Processed
5-20%
Price Improvement
03
The Meta-Aggregator: 1inch Fusion
1inch Fusion is a hybrid model that combines an RFQ system with on-chain AMM liquidity, acting as a meta-aggregator. It demonstrates the evolution from passive pools to active order flow auctions.\n- Dutch auction model: Solvers bid for order flow over time\n- Guaranteed execution: No failed transactions, users pay only on success\n- Composability: Can plug into any on-chain liquidity source via 1inch Aggregation
~90%
Fill Rate
$1B+
Fusion Volume
04
The Cross-Chain Primitive: LayerZero & CCIP
Cross-chain messaging protocols like LayerZero and Chainlink CCIP are the foundational primitive for next-gen aggregation across ecosystems. They enable intent solvers to source liquidity and compute from any chain.\n- Unified liquidity layer: Solvers treat all chains as a single venue\n- Security through diversity: Different security models (oracle networks, light clients)\n- The new battleground: Aggregation is shifting from intra-chain price to inter-chain atomicity
50+
Chains Connected
$10B+
TVL Secured
05
The Endgame: Proactive Liquidity Networks
The logical conclusion is a network where liquidity is not deposited but is proactively deployed by solvers in anticipation of demand, as seen in nascent forms with DEX aggregators on Solana.\n- Liquidity follows intent: Capital is dynamically allocated to where it's needed\n- Zero idle capital: TVL becomes an irrelevant metric; velocity is king\n- Abstracted complexity: User gets the best outcome without understanding the underlying mechanics
~1000x
Capital Efficiency
Sub-second
Execution
06
The Risk: Solver Centralization
The shift to intent-based systems creates a new centralization vector: the solver network. A small group of sophisticated actors with capital and infrastructure could dominate, recreating CEX-like dynamics.\n- Barriers to entry: Requires massive capital, MEV expertise, and cross-chain infra\n- Collusion risk: Solvers could form cartels to reduce competition\n- Protocol dependency: Aggregators like CowSwap and UniswapX must actively manage solver decentralization
<10
Dominant Solvers
Critical
Protocol Risk
future-outlook
THE AGGREGATION PRIMITIVE
The Convergence: What's Next (6-24 Months)
AMMs are a primitive form of information aggregation, and their evolution will define the next generation of DeFi infrastructure.
AMMs aggregate fragmented liquidity into a single price curve, but this is a crude, capital-inefficient method. The next generation, like UniswapX and CowSwap, treats liquidity as a composable information layer, separating price discovery from execution.
Intent-based architectures are the upgrade. They replace passive liquidity pools with a network of solvers competing to fulfill user preferences, creating a competitive market for execution quality. This mirrors the evolution from on-chain order books to off-chain solver networks.
The endpoint is a universal resolver. Protocols like Across and LayerZero demonstrate the power of generalized message passing. The final state is a cross-chain intent layer where solvers source liquidity from any venue—CEXs, private OTC desks, AMM pools—transparently.
Evidence: UniswapX processed over $7B in volume in its first six months by abstracting execution. This proves demand for intent-driven, gas-agnostic trading over direct AMM interaction.
takeaways
INFORMATION AGGREGATION PRIMITIVES
TL;DR for CTOs & Architects
AMMs are not just trading venues; they are the first decentralized, on-chain mechanism for aggregating and pricing latent demand.
01
The Problem: Price Discovery is a Coordination Game
Traditional order books fail in a decentralized, high-latency environment. AMMs solve this by using a constant function market maker (CFMM) formula to provide a continuous, on-chain price feed.\n- Aggregates liquidity from passive LPs into a single, always-available counterparty.\n- Eliminates order matching latency; price is a pure function of the reserve ratio.
~500ms
Block Time
Uniswap v2
Canonical CFMM
02
The Solution: Liquidity as a Public Good
An AMM pool is a shared liquidity primitive that any application can permissionlessly query and interact with. This creates network effects far beyond a single DEX.\n- Composability: Serves as a price oracle for lending protocols like Aave and Compound.\n- Infrastructure Layer: Enables aggregators (1inch, Matcha) and intent-based systems (UniswapX, CowSwap) to source baseline liquidity.
$10B+
TVL as Utility
1000+
Integrated Protocols
03
The Limitation: Inefficient Capital & MEV
AMMs are a blunt instrument. Passive LPs suffer from impermanent loss and provide liquidity across the entire price curve, most of which is never used. This inefficiency is a tax on the system.\n- Arbitrage is required to correct prices, creating a ~$1B+ annual MEV leakage.\n- This flaw is the core driver for Concentrated Liquidity (Uniswap v3) and intent-based architectures.
-50%
LP ROI (vs. HODL)
$1B+
Annual MEV
04
The Evolution: From Primitive to Processor
Next-gen AMMs and intent-based systems treat the primitive as a settlement layer. The aggregation logic moves upstream.\n- Uniswap v4 Hooks: Allow pools to become programmable liquidity engines.\n- Intent-Based Aggregation: Protocols like Across and CowSwap solve for user intent off-chain, using AMMs only for final execution.
10x
Capital Efficiency
Uniswap v4
Programmable
ENQUIRY
Get In Touch
today.
Our experts will offer a free quote and a 30min call to discuss your project.
NDA Protected
Confidentiality24h Response
Time to ValueDirectly to Engineering Team
No Sales Fluff10+
Protocols Shipped
$20M+
TVL Overall