Case Study · EigenLayer Season 2

$2.17M in EIGEN Tokens
Gone in 24 Hours

EigenLayer airdropped tokens to 11,176 wallets. 63% immediately dumped them on centralized exchanges. A single behavioral filter could have prevented 76% of the leakage.

63%Dumped7,042 of 11,176 wallets
68.4%Sold within 24 hours4,815 of 7,042 dumpers
2.17MEIGEN leakedTokens sent to CEX exchanges

How value leaked

1

EigenLayer

Protocol distributes 3.32M EIGEN tokens

2

Airdrop

11,176 wallets receive tokens

3

63% Dump

7,042 wallets sell within days

4

CEX Exchanges

2.17M EIGEN liquidated instantly

The Problem

Tokens Meant for Builders
Went to Extractors

EigenLayer designed Season 2 to reward genuine restakers — users who locked capital and contributed to network security.

What actually happened: 81.1% of recipient wallets (9,068 of 11,176) had behavioral scores below 200 out of 1,000. These low-quality wallets extracted 1.66 million EIGEN and dumped it on exchanges — median time to sell: 0 days.

The airdrop didn't just miss its target — it actively funded the wallets that extract value from every protocol they touch.

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

  • $20/month AI agents can farm dozens of protocols simultaneously at machine speed.
  • They simulate organic behavior, learn from each airdrop's criteria, and adapt faster than any manual filter.
  • Without behavioral intelligence that looks at patterns the next wave of leakage will be hundreds of millions.
The Evidence

Three Behavioral Patterns
Drove 63% Leakage

Analysis of 11,176 wallet records reveals clear, measurable patterns distinguishing extractors from genuine participants.

68.4%Instant Exits4,815 of 7,042 dumpers

of bad wallets sold within 24 hours. 91.2% sold within 30 days. Median time to dump: same day.

81.1%Low-Score Concentration9,068 of 11,176 wallets

of all recipients scored below 200 out of 1,000. These wallets accounted for 65.3% of all dumps.

28.7%Repeat Offenders2,022 of 7,042 dumpers

of dumpers had received and sold previous airdrops too. Bad wallets also had younger accounts — 1.12 years median vs 1.50 years for good wallets.

Wallet Outcomes by Score Range

Higher scores correlate with better behavior

Token Dumping Patterns by zScore

Historical bad behavior vs EIGEN airdrop dumping by score range

The Solution

If zScore Was Used

zScore was not live during EigenLayer Season 2. The analysis below simulates outcomes if a behavioral filter (zScore ≥ 200) had been applied to the observed dataset — a conservative threshold excluding only the lowest-quality profiles.

The Bottom Line

76%Leakage PreventedToken waste reduction
1.66MEIGEN SavedTokens kept from dumpers
5,919Bad Wallets BlockedExtractors filtered out

One behavioral filter — excluding wallets scoring below 200 — would have prevented three-quarters of all token leakage, saving 1.66 million EIGEN from immediate liquidation.

MetricActual (No Filter)With zScore ≥ 200Impact
Eligible wallets11,1762,108−81%
Bad wallets7,042 (63.0%)1,123 (53.3%)−84%
EIGEN leaked2,174,572517,469−76%
EIGEN saved1,657,103
01

Set Filter Thresholds

Start with zScore ≥ 200 as baseline eligibility. Protocols can tune thresholds based on their risk tolerance and community size.

02

Monitor for Gaming

Track score distribution shifts before and during claim windows. Flag sudden score increases or cluster patterns that suggest coordinated manipulation.

03

Shadow → Enforce → Iterate

Run filters in shadow mode for one cycle. Compare actual vs filtered outcomes. Enforce in the next cycle. Iterate thresholds based on observed results.

What Comes Next

For Protocols

  • 01

    Score before you distribute. Use behavioral data to assess wallet quality before committing tokens.

  • 02

    Implement graduated vesting. Tie release schedules to post-airdrop behavior, not just eligibility.

  • 03

    Monitor and iterate. Run filters in shadow mode first. Measure false positive rates. Tune thresholds.

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 filter airdrop recipients, underwrite credit risk, assess agent reliability, and allocate incentives based on real contribution — not just eligibility checklists.

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