The logs show 85% idle capital across seven chains.
Not a typo. Not a single bad pool. This is the aggregate state of concentrated liquidity market making (CLMM) on the largest DEX forks. 1inch commissioned Dune Analytics to run the numbers. The result is a forensic indictment of how we deploy capital in DeFi. The code did not lie; the humans misread the data.
Context: The CLMM Promise and Its Unseen Cost
Uniswap V3 introduced a paradigm shift. Instead of providing liquidity across the full price curve (0 to infinity), LPs could concentrate their capital within a specific range. In theory, this should dramatically improve capital efficiency. The numbers, however, tell a different story. The study, based on 2026 H1 on-chain data, scanned over 10 million position records across Ethereum, Arbitrum, Optimism, Base, Polygon, zkSync Era, and Avalanche. The methodology was simple: measure the overlap between an LP's provided price range and the actual active trading range over a rolling 30-day window. If 100% of the range is active, utilization is 100%. If only the bottom 10% of the range touches price, utilization is ~10%. The aggregate finding? 85% of all CLMM liquidity is structurally underutilized.
Core: The On-Chain Evidence Chain
Let's break down the signal from the noise. The headline figure is dramatic, but the forensic breakdown is more revealing.
First, the outlier correction. The 85% figure is a weighted average. When you remove the top 1% of positions by capital (whales and professional market makers), the underutilization rate jumps to 91%. This tells me the small-to-medium LP is the primary victim. Professional firms use algorithms to dynamically rebalance their ranges. Retail LPs set and forget. The gap is existential.
Second, the out-of-range problem. 29.5% of all capital locked in CLMM pools is completely outside the active price range. This means millions of dollars sit in “zombie” positions—earning zero fees, exposed to impermanent loss, and serving no function to the market. If we extrapolate this to the $170 billion in total CLMM TVL across these chains, roughly $50 billion is purely decorative. It provides no utility.
Third, the chain-by-chain breakdown. Ethereum Mainnet shows the highest absolute idle capital (~$12B). But the percentage is highest on Base and zkSync Era, both exceeding 90%. This aligns with my prior experience analyzing Arbitrum’s TVL decay. Newer chains attract speculative LPs who deploy wide ranges out of fear of being left behind. They do not pay the management cost. The data confirms this.
Fourth, the pool-level dispersion. The issue is not uniform. Pools with high natural volatility (e.g., ETH/BTC) see 70-80% utilization. Stablecoin pairs (USDC/USDT) surprisingly show only 60% utilization. The worst performers are mid-cap altcoin pairs, where the implied volatility is high but the actual trading volume is low. LPs overestimate the need for massive range to capture fees that never materialize. This is a classic capital misallocation trap.
Fifth, the time decay analysis. The study also sliced data by position age. Positions open for less than 30 days have a 75% underutilization rate. Positions open for more than 180 days? 92%. This is a clear signal of LP attrition over time. The longer you hold a static CLMM position, the more likely it decays into irrelevance. Transition is not an event, but a data stream.
The $1.7 Billion Opportunity Cost
To quantify the damage, I built a simple model based on the study's findings. If the idle 85% could be redeployed at the average CLMM fee rate (0.05% per trade, assuming 1x turnover per day), the annualized opportunity cost across these seven chains exceeds $1.7 billion. This is not a small optimization. This is a major structural inefficiency that acts as a tax on passive LPs.
Contrarian Angle: Correlation Is Not Causation
Before we declare CLMM a failed experiment, let me offer the counter-read. The study treats all “underutilization” as waste. But professional market makers will argue that defensive liquidity is rational. In a volatile market, you maintain a wider range to avoid being knocked out by a sudden spike. The 85% figure includes this strategic slack.
Counter-evidence #1: The study does not distinguish between “intentional” wide-range strategies (used by sophisticated MM firms to manage risk) and “accidental” wide-range strategies used by retail. The aggregate data obscures this. If 20% of the idle capital is intentional, the waste figure drops to 68%—still massive, but less apocalyptic.
Counter-evidence #2: The study uses a 30-day window. Under extreme volatility, the active range can shift by 50% within a week. A position that appears “idle” for the full month might have been perfectly productive for the first 15 days. The metric is mean-sensitive.
Counter-evidence #3: The study does not consider the opportunity cost of rebalancing. To keep a position constantly in-range, LPs must pay gas fees (on L1) or management fees (on L2). The 85% idle figure assumes zero cost to maintain high utilization. In reality, the optimal utilization rate is not 100%. There is a trade-off between capital efficiency and transaction cost.
My verdict: The direction is correct. The magnitude is directionally accurate, but the headline number should be interpreted as “85% is less efficient than optimal,” not “85% is wasted.” The true waste is likely between 60-70% for retail LPs. This is still a scandal.
Takeaway: The Signal for Next Week
This report is not a post-mortem. It is a product roadmap. The implications are clear:
- Aggregators win. 1inch’s business model—finding optimal execution paths across fragmented liquidity—is validated. If DEXs are structurally inefficient, the middleware that filters them becomes indispensable.
- Active management protocols benefit. Protocols like Arrakis, Mellow, and even automated vaults on Maverick will see capital inflows. The era of “set and forget” CLMM is ending.
- Uniswap V3 faces pressure. The 85% number is a direct challenge to its core value proposition. V4’s “hooks” are an attempt to solve this, but the complexity spike is real.
Final question: Will the market reward the diagnostic (1inch, Dune) or the cure (active management protocols)? I am short on human patience and long on automated execution. The data does not change behavior; it just reveals the cost of inertia. The market will correct this inefficiency.
