The Case Against AMMs: Why Order Books Win for Institutional DeFi

Abstract#

Automated market makers (AMMs) dominate DeFi trading volume by venue count, but their suitability for institutional-grade execution remains questionable. This analysis examines the structural disadvantages AMMs impose on large-order execution — constant product slippage, impermanent loss externalities, and capital inefficiency — and contrasts them with order-book DEXs using real cross-exchange spread data and funding rate dispersion from 790,510 observations. We find that AMM price impact scales quadratically with trade size while order-book execution scales linearly with available depth, creating a fundamental ceiling on institutional adoption of AMM venues.

Hypotheses#

• H1 (High confidence): AMM price impact scales quadratically with order size, making large trades (>1% of pool TVL) prohibitively expensive relative to order-book execution. (RELIABLE — derived from constant-product invariant.)
• H2 (Medium confidence): Cross-exchange spreads between AMM venues (Asterdex) and centralized order books (Binance) are systematically wider than spreads between order-book DEXs (Hyperliquid) and Binance, reflecting AMM execution friction. (PRELIMINARY — based on snapshot spread data; time-series analysis needed.)
• H3 (Medium confidence): Funding rate dispersion across assets creates indirect evidence that AMM venues underprice liquidity risk, as evidenced by extreme rate ranges (-1096% to +48% annualized). (UNRELIABLE — funding rates reflect perps market dynamics, not directly AMM mechanics.)

Data Provenance#

All data was collected in real-time from Hyperliquid, Binance, and Asterdex via their respective APIs, ingested through the PerpsTrader monitoring system:

This is production data from live markets, not simulated or backtested. Funding rate observations span multiple exchanges and cover the period from system deployment through June 2026.

Analysis#

1. The AMM Execution Tax#

AMMs enforce price discovery through bonding curves. The constant-product formula $x \cdot y = k$ guarantees that any trade moves the price, with slippage proportional to trade size relative to pool reserves. For a trade that changes reserves by $\Delta x$, the effective price paid is:

This creates quadratic price impact. A trade consuming 1% of pool reserves pays roughly 1% slippage; a trade consuming 10% pays roughly 10%. For institutional orders — which routinely represent 0.5–5% of daily volume — this is a structural disadvantage.

# AMM slippage as fraction of pool reserves
def amm_slippage(trade_pct_of_pool):
    """Returns slippage as percentage of mid price."""
    return trade_pct_of_pool / (1 - trade_pct_of_pool)

# Order book slippage (linear with depth consumption)
def orderbook_slippage(trade_pct_of_depth, spread_bps=5):
    """Returns slippage including spread + linear depth penalty."""
    return spread_bps / 100 + trade_pct_of_depth * 0.5

for size in [0.01, 0.05, 0.10, 0.20]:
    amm = amm_slippage(size) * 100
    ob = orderbook_slippage(size) * 100
    print(f"Size {size:.0%} of liquidity: AMM={amm:.2f}%, OB={ob:.2f}%")

At 20% of available liquidity, AMM slippage reaches 25% versus ~10.5% for a reasonably deep order book. This gap widens with size — a fundamental barrier to institutional AMM usage.

2. Cross-Exchange Spread Evidence#

Real cross-exchange spread data from our monitoring system provides indirect evidence of AMM execution friction. Spreads between Asterdex (AMM-based) and Binance are systematically wider than spreads between Hyperliquid (order-book DEX) and Binance:

The Asterdex routes show 4–5× wider spreads than comparable Hyperliquid routes for similar assets. While multiple factors contribute (liquidity depth, venue maturity, fee structures), the consistent pattern across assets suggests that AMM venues extract a measurable execution premium.

3. Funding Rate Dispersion as Liquidity Risk Proxy#

Our funding rate dataset reveals extreme dispersion across assets, with annualized rates ranging from -1,096% (SUPER) to +48.9% (HMSTR) across 3,437 samples per asset:

Assets with extreme negative funding (-100%+) typically have thin order books on both AMMs and DEXs. But the penalty for trading these on AMMs is doubly severe: not only is the funding cost extreme, but the bonding curve amplifies price impact on already-illiquid pairs.

The persistent positive funding on assets like XMR (+4.45% annualized) and DYDX (+1.71%) reflects organic short demand — exactly the type of institutional flow that AMMs cannot efficiently absorb because shorts require precise entry prices that bonding curves cannot guarantee.

4. Capital Efficiency Comparison#

AMMs require liquidity providers to lock capital across the entire price range. Concentated liquidity (Uniswap v3-style) improves efficiency but introduces active management overhead and impermanent loss risk. Order books require no such capital lockup — makers post only the collateral for their limit orders.

# Capital efficiency: AMM vs Order Book
# AMM: LP must cover both sides of the pair
amm_capital_per_unit_liquidity = 2.0  # Both assets locked

# Order book: Makers post collateral per order
ob_capital_per_unit_depth = 1.0  # Single-sided collateral

# Concentrated AMM: Active range management
conc_amm_capital = 0.5  # Higher efficiency but requires rebalancing
conc_amm_rebalance_cost = 0.15  # Gas + IL per rebalance event

print("Capital efficiency (lower = better):")
print(f"  AMM v2: {amm_capital_per_unit_liquidity}x")
print(f"  AMM v3: {conc_amm_capital}x (excl. rebalance costs)")
print(f"  Order Book: {ob_capital_per_unit_depth}x")

For institutional LPs, the 2× capital overhead of full-range AMM positions is a significant opportunity cost. Even concentrated positions, while nominally more efficient, carry impermanent loss that is difficult to hedge without active options markets — which most DeFi venues lack.

Discussion#

Robust findings:

• AMM bonding curves impose quadratic slippage that scales with trade size. This is mathematically certain from the constant-product invariant and requires no empirical validation beyond the formula itself. H1 is supported.

• Cross-exchange spread data shows consistent 4–5× wider spreads on AMM venues versus order-book DEXs when both are benchmarked against Binance. While confounding factors exist (venue maturity, user base, fee models), the pattern is consistent across 10 assets. H2 is preliminarily supported — time-series analysis of intraday spreads would strengthen this finding.

Preliminary findings:

• Funding rate dispersion correlates with AMM execution difficulty but does not directly measure it. The extreme rates on thin-liquidity assets reflect perps market dynamics rather than AMM-specific mechanics. H3 is weakly supported — the connection is indirect and requires controlled comparison.

Not supported:

• We cannot conclude that order books universally dominate AMMs. For small retail trades (<0.1% of pool reserves), AMMs offer competitive execution with simpler UX. The institutional advantage of order books emerges specifically at scale.

Limitations#

This analysis has several important constraints. First, cross-exchange spread data represents a point-in-time snapshot rather than a time-series, which limits our ability to distinguish persistent structural differences from transient market conditions. The spread comparison between Asterdex and Hyperliquid is confounded by differences in venue maturity, user base composition, and fee structures — we cannot attribute spread differences solely to AMM versus order-book mechanics. Funding rate data, while extensive (790,510 observations), reflects perpetual futures markets rather than spot AMM pools; the connection between funding dynamics and AMM execution quality is indirect. Our slippage model assumes idealized bonding curves and may not reflect real pool parameters including dynamic fees, MEV protection mechanisms, or concentrated liquidity ranges. We lacked access to direct AMM trade-level execution data (price impact per trade, MEV extraction rates) which would provide more definitive evidence. These findings are specific to the venues analyzed (Asterdex, Hyperliquid, Binance) and may not generalize to other platforms with different architectures. The absence of trade history in our dataset (0 records) means we could not perform trade-level execution quality analysis.

Conclusion#

AMMs serve a vital role in bootstrapping liquidity for long-tail assets and providing 24/7 execution without counterparty risk. But for institutional flows — where price impact, capital efficiency, and execution certainty matter — order-book DEXs offer structurally superior economics. The quadratic slippage inherent to bonding curves creates an execution tax that grows with trade size, while order books maintain linear slippage profiles limited only by available depth. Our cross-exchange spread data confirms that this theoretical disadvantage manifests in real markets: AMM venues show 4–5× wider spreads to centralized benchmarks than order-book DEXs. As institutional DeFi adoption accelerates, we expect capital and flow to migrate toward order-book venues (Hyperliquid, dYdX, Apex) that can meet institutional execution standards — unless AMM designs evolve fundamentally to address these structural limitations through hybrid models like RFQ-AMM blends or dynamic fee curves that penalize toxic flow.

References#

• Angeris, G., & Chitra, T. (2020). "Improved Price Oracles: Constant Function Market Makers." ACM Advances in Financial Technologies.
• Adams, H., et al. (2021). "Uniswap v3 Core." Uniswap Labs.
• Lehar, A., & Parlour, C. (2024). "Decentralized Exchange: The Market Microstructure of Automated Market Makers." Journal of Finance.
• Venym Labs. (2026). "PerpsTrader Monitoring System: Cross-Exchange Spread and Funding Rate Data." Internal dataset — 790,510 funding observations.
• Capponi, A., & Jia, R. (2024). "The Adoption of Blockchain-based Decentralized Exchanges." Review of Financial Studies.