I ran a deterministic EVM simulation of 159 active Aave V3 positions on Ethereum mainnet across four ETH price shock scenarios. This post shares the methodology, findings, and a full PDF report.
Motivation
Aave is at an unusual inflection point. V4 has just launched, core risk contributors are in transition, and V3 will require active oversight for months — potentially years — until migration is complete. Understanding the current liquidation exposure on V3 is directly relevant to whoever takes on risk coverage during this period.
This analysis provides a transaction-level view of cascade risk under ETH price declines. It is not a statistical model or a probabilistic estimate — it is the exact health factor computation Aave’s liquidation logic would perform on-chain.
Methodology
The simulation forks live Ethereum mainnet state at a specific block using REVM — the Rust EVM implementation used by Reth and Foundry. Price shocks are applied by replacing Chainlink aggregator bytecode in the forked state with mock contracts returning the specified price. Volatile assets (WETH, WBTC, wstETH, cbETH, rETH) are shocked proportionally. Stablecoins hold their $1.00 peg.
getUserAccountData() is then executed against this modified state. The output is deterministic and reproducible — the same values Aave’s contracts would compute under those exact conditions.
The 159 positions analysed are a filtered subset (HF < 1.3) sourced from The Graph Aave V3 subgraph. Position coverage is being expanded to capture larger whale positions outside the subgraph borrow event set.
Findings
| ETH Price | Drop | Liquidatable Positions | Collateral at Risk |
|---|---|---|---|
| $2,000 | -10% | 10 | $3.1M |
| $1,500 | -33% | 56 | $25.4M |
| $1,200 | -46% | 75 | $126.9M |
| $1,000 | -55% | 80 | $127.9M |
At current prices (~$2,231) no positions in the analysed set are immediately liquidatable.
Key Observations
The sharpest liquidation cliff sits between $2,000 and $1,500 ETH — positions jump from 10 to 56 liquidatable in that range, with collateral at risk rising from $3.1M to $25.4M. This suggests a cluster of positions holding health factors in the 1.05–1.20 band with concentrated ETH exposure.
A 33% single-move ETH decline is within historical precedent. During the V3-to-V4 migration window, with liquidity potentially fragmented across both deployments, cascade response times may be slower than under stable conditions. This window warrants closer monitoring than normal.
Full Report
[phantom_report.pdf] — 8-page report with per-position health factor breakdown across all four scenarios, methodology detail, and cascade risk charts.
https://github.com/sitwiz/phantom-sim/releases/download/v0.1.0/phantom_report.pdf
Happy to run this against specific position subsets, additional price scenarios, or targeted whale positions if useful to the community or incoming risk service providers.