Aave V4 shipped this week, delivering its long-awaited unified-liquidity architecture across the chains it operates on. First-72-hour data: $1.6B of V3 TVL migrated to V4, supply rates compressed 12% on USDC and USDT pools, and gas costs for cross-chain borrowing fell roughly 40%. It is early — but the bet that lending becomes a multi-chain product, not a per-chain product, is now in the live market.
Key takeaways
- V4 unified-liquidity TVL: $1.6B in first 72 hours (out of ~$13B total Aave TVL).
- USDC + USDT supply APY compressed 8–12% as deeper pools reduced rate volatility.
- Cross-chain borrowing gas cost down ~40% on Optimism→Arbitrum flows.
- V4 introduces a Hub-and-Spoke architecture: liquidity lives in Hubs (per chain), markets are Spokes (per asset).
- Aave (AAVE) token model unchanged in V4; revenue accrues the same way.
What V4 actually changed
The V3 design held liquidity in per-asset pools on each chain. If you wanted to borrow USDC on Arbitrum, you tapped the Arbitrum USDC pool. If that pool was thin, your rate spiked. The result was rate fragmentation — the same asset traded at different rates on different chains, and large borrowers had to chain-hop to find depth.
V4 introduces a “Hub” abstraction. Liquidity is consolidated into per-chain Hubs. Asset markets (Spokes) tap the Hub. The result: deeper effective pools, less rate spikes, easier risk management for the protocol.
The 72-hour data
Migration is staged — Aave Labs released a migration tool that V3 LPs can opt into. In the first 72 hours, $1.6B of TVL crossed over (12% of total Aave TVL). The migration leader was USDC (43% of migrated volume), followed by USDT (28%) and ETH (14%).
The clearest market-level effect: supply rates on USDC and USDT compressed by 8–12% as pool depth increased. Borrow rates fell roughly proportionally. The net spread (which Aave captures) stayed flat.
Cross-chain UX
The bigger story is cross-chain. V4’s design lets a borrower on Optimism collateralise with assets sitting in the Arbitrum Hub — the unified Hub-Spoke layer handles the routing. Initial measurements show cross-chain borrowing gas costs down ~40% vs the equivalent V3 manual flow (bridge → deposit → borrow). Time-to-completion dropped from ~5 minutes to ~90 seconds.
For institutional borrowers running treasury operations across multiple chains, this is the headline feature. Whether it converts retail behaviour is a separate question — most retail users borrow on the chain they already sit on.
What it means for AAVE token holders
The token model in V4 is largely unchanged. AAVE captures revenue via the staking module (sAAVE), and the revenue formula is the same: a slice of borrow-rate spread accrues to the safety module. Deeper pools mean higher absolute volume, which mathematically should mean higher revenue. The compressed spread per dollar is roughly offset by larger dollar volume.
The execution risk is the migration itself — V3 keeps running in parallel for 18 months. If migration stalls, the unified-liquidity thesis weakens. Worth watching.
“V4 turns lending from a per-chain product into a credit primitive that happens to live on multiple chains. That’s a different business.”
| Metric | V3 (pre-launch) | V4 (72h in) |
|---|---|---|
| USDC supply APY | 5.20% | 4.58% |
| USDT supply APY | 5.41% | 4.79% |
| ETH supply APY | 2.04% | 2.01% |
| Cross-chain borrow gas | ~$8.20 | ~$5.05 |
| Cross-chain borrow time | ~5 min | ~90 sec |
Three numbers to watch this week
- Migration pace — does $1.6B in 72h continue at $500M/day, or does it stall?
- V3 vs V4 cross-rate — if the same asset trades at different rates on V3 vs V4, arbitrageurs will close it fast; if not, the migration design has a flaw.
- Oracle liveness — V4 adds new oracle paths; an oracle failure in the first month would be a major reputational hit.
Why this matters
Aave is the largest DeFi protocol by TVL. Whatever architectural decisions it makes propagate across DeFi within months. V4’s unified-liquidity bet, if it works, becomes the template other money markets copy. If it doesn’t — if migration stalls or an oracle failure exposes a Hub-level vulnerability — the per-chain pool design returns as the default.
The early data is mildly bullish. Check the live AAVE page for the model’s prediction.