Aave V4 mainnet deployment completed in November 2025. Six months in, the realized TVL migration from V3 to V4 sits at approximately $4.8 billion of total V4 TVL versus approximately $11.2 billion remaining on V3 as of late April 2026. The combined Aave protocol TVL of approximately $16 billion is broadly stable relative to the pre-V4-deployment baseline; what has changed is the realized split between the two versions, which reflects how lenders and borrowers have actually adopted the V4 framework over the operational window. I have been pulling the V3-versus-V4 deposit and borrow data into my workbench through Q1 and Q2 and the realized migration pattern is structurally informative about how DeFi protocol upgrades actually translate into user adoption.
The structural fact that anchors the analysis: V4's adoption rate has been slower than the V2-to-V3 migration rate (which completed approximately 80% of TVL migration within 4-6 months of V3 deployment). The realized V4 adoption shows approximately 30% of total Aave TVL has migrated within 6 months — a meaningfully slower pace that reflects specific structural reasons users have for staying on V3.
The Realized V4 Architectural Differences That Matter Operationally
Before the realized data, a quick framing of what V4 actually changes that affects user adoption decisions. V4 introduced (1) a unified liquidity layer across multiple borrow markets that improves capital efficiency for lenders, (2) a new GHO stablecoin integration that enables certain borrow strategies that V3 did not accommodate, (3) revised interest rate models with realized rate calculation that responds more dynamically to utilization changes, and (4) new collateral categories including LRT collateral and certain RWA collateral types that V3 did not support.
For most retail and mid-tier DeFi users, these architectural changes do not translate into materially different user-facing experience. The realized borrow and lend rates on standard collateral pairs (ETH borrow against USDC, USDC borrow against ETH, stablecoin-stablecoin pairs) are within approximately 0.5-1.5 percentage points across V3 and V4 at typical utilization levels. The realized user-facing benefit of migrating from V3 to V4 for standard positioning is therefore modest.
For specific positioning that V4's architectural improvements unlock (LRT collateral borrowing, GHO-stablecoin-routed borrow strategies, capital-efficient multi-market lending), V4 provides genuine new capability. But these positioning structures are concentrated in a relatively narrow user segment.
The Q2 2026 V4 TVL Decomposition
V4 TVL of approximately $4.8 billion decomposes approximately as follows:
- Standard ETH/USDC/stablecoin lending positions (V3-equivalent positions migrated to V4): approximately $2.4 billion (50%)
- LRT collateral lending (V4-specific capability): approximately $1.2 billion (25%)
- GHO-routed borrow positions (V4-specific): approximately $0.8 billion (17%)
- Other V4-specific positioning (RWA collateral, multi-market positions): approximately $0.4 billion (8%)
The realized pattern shows approximately 50% of V4 TVL is positioning that could equivalently exist on V3 — users who migrated for non-functional reasons (interface preference, default protocol selection, signaling) rather than for V4-specific capability. The remaining 50% is V4-specific positioning that V3 cannot accommodate.
This is structurally meaningful for thinking about why V3 has retained approximately 70% of total Aave TVL through Q2 2026. Users who do not need V4-specific capability have limited reason to incur the gas cost and operational overhead of migrating positions, particularly when the realized rate differential between the two versions is modest.
The Realized Rate Spread Between V3 and V4
The interest rate model in V4 produces somewhat different equilibrium rates than V3 at equivalent utilization levels. The realized borrow and lend rates on USDC across the two versions at the late-April 2026 observation:
USDC market on V3:
- Lend APY: approximately 4.8%
- Borrow APY: approximately 6.4%
- Utilization rate: approximately 78%
- Spread: approximately 1.6 percentage points
USDC market on V4:
- Lend APY: approximately 5.2%
- Borrow APY: approximately 6.8%
- Utilization rate: approximately 81%
- Spread: approximately 1.6 percentage points
The realized rates on V4 run approximately 0.4 percentage points higher than V3 on both lend and borrow sides at typical utilization. The spread between borrow and lend rates is approximately equal across the two versions.
The structural reason for the rate differential: V4's interest rate curve responds more aggressively to utilization changes near the optimal-utilization threshold, producing somewhat higher equilibrium rates at typical utilization levels. For lenders, this is a modest positive — approximately 0.4 percentage points of additional realized yield. For borrowers, it is a corresponding modest negative — approximately 0.4 percentage points of additional realized borrowing cost.
Why Most V3 Users Have Not Migrated
Three structural reasons V3 users have stayed on V3 despite V4's modest rate advantages.
First, the gas cost of migration. Migrating a non-trivial position from V3 to V4 requires multiple onchain transactions — closing the V3 position, transferring assets, opening the equivalent V4 position. At typical Ethereum mainnet gas prices, the realized migration cost runs approximately $80-180 in gas for a standard position. For a position size of approximately $5,000-$50,000, this gas cost represents approximately 0.16-3.6% of position value — meaningful relative to the approximately 0.4 percentage point annualized rate advantage that migration captures.
Second, the operational risk of position recreation. Closing and reopening positions introduces realized risk of execution errors, position-state changes during the transition window, and operational mistakes that produce realized loss. Users who are comfortable with their V3 positioning and do not need V4-specific capability have limited reason to incur this operational risk.
Third, the social and informational cost of migration. Most DeFi users who hold V3 positions had to learn V3's specific interface, parameters, and operational quirks. Migrating to V4 requires learning the (somewhat different) V4 interface and operational characteristics. For users who treat their DeFi positioning as relatively passive, this learning cost outweighs the modest rate benefit.
What V4 Has Actually Delivered Where It Matters
The V4-specific capability delivery has been meaningful where it matters. Three specific use cases that V4 has unlocked in operational terms.
LRT collateral borrow positions. V4's LRT collateral integration has enabled approximately $1.2 billion of borrow positioning that V3 could not accommodate. Users borrowing stablecoins against weETH, eETH, and rsETH collateral can now do so on Aave directly rather than through alternative protocols. The realized capital efficiency gain for these users is meaningful — approximately 65-75% LTV on LRT collateral on V4 versus approximately 0% (no LRT collateral support) on V3.
GHO-routed positions. V4's GHO stablecoin integration has enabled approximately $0.8 billion of borrow positioning that benefits from Aave-protocol-native GHO economics. Users borrowing GHO against ETH or stablecoin collateral on V4 face slightly different cost economics than borrowing standard stablecoins, with the realized differential favoring GHO under specific positioning conditions.
Capital-efficient multi-market positioning. V4's unified liquidity layer enables specific multi-market positioning structures that capture cross-market rate inefficiencies. The realized capital efficiency for these positions is approximately 15-25% better than equivalent positions split across V3's separate market structures. The user segment running these structures is small but the realized capability is genuinely new.
My Read On Aave's Multi-Version Operating Model
The Aave protocol's realized operating model with V3 and V4 in parallel has been structurally interesting. Rather than forcing migration, the protocol has accepted that V3 and V4 will continue operating in parallel for an extended period — possibly years — with users migrating only when their specific positioning benefits from V4's capability.
This is structurally different from how Compound and other lending protocols have managed version transitions. Compound's V2-to-V3 transition accelerated through deprecation of V2 functionality; Aave's V3-to-V4 transition is more gradual and respects user preference for staying on V3 if their positioning works there.
For traders evaluating their own DeFi lending positioning, the structural read is that V3 remains a fully operational protocol with no near-term deprecation risk. Users who have V3 positions that work for their needs do not need to migrate to V4 unless V4-specific capability matters for their use case.
Honest Limits
I did not run direct contract-level audits of Aave V3 or V4 — the TVL and rate figures referenced here come from publicly disclosed Aave protocol data through DeFi Llama and protocol-direct dashboards, not granular contract-level reconstruction. The V4 TVL decomposition by use case reflects approximate categorization from realized borrow and lend pattern analysis and may not capture every position type precisely. The realized rate differential between V3 and V4 reflects late-April 2026 observation and may shift as utilization patterns evolve. The user-segment behavioral inference about why most V3 users have not migrated reflects my reading of the realized pattern rather than direct user research. The personal positioning observations reflect my own DeFi positioning workbench and are not investment advice or recommended allocation. The Aave protocol's multi-version operating model may shift through 2026 if governance decisions or operational conditions change the deprecation timeline for V3.