Research Report · Aave V3 Liquidity Providers

4.9% of Wallets Hold 93% of TVL
And Protocols Still Reward Them Equally

Aave V3 on Ethereum has 132,064 liquidity providers. A tiny cohort of 6,460 high-zScore wallets delivers nearly all the protocol's depth. Flat incentive structures reward everyone equally — including wallets that contribute nothing.

93%of TVL from 4.9% of wallets6,460 wallets in the 900–1000 band
241,129×Value-per-gas efficiencyTop wallets vs. low-score LPs
3.1 yrsAvg tenure of core LPs1,129 active days in 900–1000 band

How incentives are currently distributed

1

Aave V3

Protocol emits uniform incentives

2

132,064 LPs

All wallets treated as equal recipients

3

95.1% Low-Impact

125,604 wallets with near-zero TVL contribution

4

Wasted Spend

Incentives absorbed by passive participants

The Problem

DeFi Protocols Reward All LPs the Same
Regardless of How Much They Actually Contribute

Aave V3 depends on a deep, reliable pool of liquidity providers to function. Protocol incentives — emissions, fee sharing, governance weight — are distributed to attract and retain the wallets that supply this depth.

What actually happens: 9,633 wallets in the 0–100 band contribute exactly $0 in average TVL — inactive or dust wallets absorbing the same eligibility treatment as whale LPs who average $1.7B per wallet and have been active for over 3 years.

A protocol cannot grow by rewarding presence. It grows by rewarding contribution.

In the Age of AI Agents, This Gets 100× Worse

  • $20/month AI agents can spin up hundreds of LP positions with minimal capital — qualifying for the same incentive tiers as wallets managing hundreds of millions.
  • They deposit dust amounts, hold through snapshot windows, and withdraw immediately after — manufacturing the appearance of sustained liquidity provision while farming rewards at near-zero cost.
  • Without behavioral intelligence that measures actual contribution depth, the next incentive round will pay low-quality wallets at industrial scale.
The Evidence

Three Patterns Prove
zScore Predicts LP Quality With Precision

Analysis of 132,064 Aave V3 LP wallets shows that zScore is not just correlated with contribution quality — it predicts it with a precision that no activity-count metric achieves.

93%Extreme TVL Concentration6,460 of 132,064 wallets

of all Aave V3 TVL is held by wallets scoring 900–1000. These 6,460 wallets maintain an average of $1.7B each. The bottom 53% of wallets by count contribute less than 0.01% of total protocol liquidity.

241K×Exponential Capital EfficiencyValue-per-gas ratio at 900–1000

value-per-gas advantage for top-band wallets over the lowest active band. This ratio compounds exponentially with zScore — a signal so strong it cannot be faked with superficial transaction volume alone.

3.1 yrsCommitment Gradient1,129 active days in top band

average tenure for wallets in the 900–1000 band — versus 1 day for the lowest band. The highest LPs engage across 42 protocols on average, demonstrating deep ecosystem integration, not isolated farming.

LP Wallet Distribution by zScore

Most wallets cluster in mid-to-high bands — yet TVL lives at the top

Wallet Share vs. TVL Share

4.9% of wallets control 93% of liquidity — the incentive allocation mismatch

The Solution

Score-Weighted Incentives

The goal is not to exclude low-score wallets — it is to allocate incentive weight in proportion to behavioral contribution. zScore provides a reliable, manipulation-resistant basis for doing so.

The Opportunity

93%TVL ConcentrationLives in 4.9% of wallets
241K×Capital Efficiency GapTop vs. bottom LP cohort
25×Incentive MultiplierFor top-band vs. passive LPs

Redirecting just 10% of flat incentive spend toward wallets scoring above 700 would concentrate rewards on the cohort responsible for 99.9% of protocol liquidity — without excluding any wallet from participation.

zScore BandWalletsAvg TVL / WalletAvg Active TenureSuggested Weight
0–1009,633$01 dayBase only
400–60026,415$316 – $9,96797 – 222 days1× – 2×
600–70032,967$213,694379 days
700–80030,881$2.9M646 days5× – 8×
800–90024,365$29.8M1,048 days10× – 15×
900–10006,460$1.7B1,129 days20× – 25×
01

Segment Before You Distribute

Score every LP wallet before the next incentive snapshot. A single zScore API call per wallet is sufficient to assign a distribution tier — no new infrastructure required, no changes to existing smart contracts.

02

Apply Multipliers, Not Cutoffs

Do not block low-score wallets from incentives — reduce their weight proportionally. Even a 1× baseline for 0–400 wallets, scaling to 25× for 900–1000, dramatically shifts effective emissions toward the wallets that sustain the protocol.

03

Measure → Iterate → Publish

After one incentive cycle, compare TVL per incentive dollar across tiers. Publish the results. Transparent, data-driven allocation builds trust with serious LPs — the ones who decide where to park hundreds of millions.

What Comes Next

For Protocols

  • 01

    Score your LPs before the next incentive round. 93% of your TVL comes from 4.9% of wallets. If your incentive structure does not reflect this, you are subsidizing passive participants at the expense of your core contributors.

  • 02

    Replace flat emissions with contribution-weighted distributions. Use zScore tiers as multipliers — this does not require new smart contracts, just a pre-snapshot segmentation call on existing wallet lists.

  • 03

    Retain your core LP cohort deliberately. Wallets in the 800–1000 band have been active for 1,000+ days across 24–42 protocols. They are committed ecosystem participants. They respond to recognition and proportional reward — not flat incentives.

What ZeruAI Provides

zScore — a universal behavioral reputation score (0–1,000) for every EVM and non-EVM wallet, derived from onchain activity across 40+ chains. Queryable via API. Mintable as an onchain credential.

Protocols integrate zScore to weight LP incentives by contribution quality, identify core liquidity providers, filter airdrop recipients, underwrite credit risk, and allocate resources based on real behavioral depth — not wallet counts or raw transaction volume.

Every number in this report is derived from public onchain data. Download the raw dataset to verify the findings independently.