Interpreting Aave V4: A Transition from Product to "Bank"

Author: Eric, Foresight News

Late on March 30, Beijing time, the Aave V4 version—which was approved for development starting in 2024—was officially launched on the mainnet, bringing the first good news since the Aave DAO governance controversy.

The V4 version can be described as a complete overhaul and rebuild of Aave. The most core change is that the previously independent lending markets have been integrated into a unified liquidity pool architecture: Hub and Spoke.

In the V4 version, each chain or L2 has a unified liquidity hub (the Hub). All assets deposited by users for lending are stored in a single liquidity pool. The Hub is responsible for global coordination, credit limit control, system-level constraints (such as “total borrowings ≤ total supplied assets”), and emergency pauses. The Hub does not directly interface with users; it manages liquidity centrally in the background.

It’s worth noting that on each chain there isn’t only one Hub; instead, different Hubs are designed based on specific needs. Essentially, this is also a form of risk isolation. For example, V4 currently has launched the Core Hub, Prime Hub, and Plus Hub. The Core Hub includes mainstream assets and is accessible to all users. The Prime Hub is designed for suppliers seeking more “controllable” collateral. The Plus Hub is tailored for strategy-based stablecoins, with parameters designed considering the project’s scale.

As for the Spoke, you can think of it as an independent market. Each market has its own lending and borrowing functions, risk parameters, and collateral rules. Within a Hub, users’ assets are stored in the same liquidity pool, and borrowers need to select different Spokes based on their needs. For example, as shown in the diagram above, users can deposit WETH as a borrowable asset. Borrowers can borrow WETH from the first four Spokes, but only the EtherFi Spoke allows collateralization with weETH.

While the official claim is that it can consolidate fragmented liquidity, in practice, for users borrowing against high-quality collateral assets, the difference isn’t significant. For example, if you want to collateralize ETH to borrow, there’s no difference in the process between V3 and V4—so long as you can ensure the health factor doesn’t drop too low.

Therefore, in terms of liquidity integration, V4 is indeed more refined than managing independent markets, but it’s not a qualitative leap. The real innovations are the customized parameters for each Spoke and the new liquidation engine.

In V4, the borrower’s interest rate depends on the base rate and the risk premium. The base rate still uses the utilization curve similar to V3—rising slowly below the optimal utilization, and increasing sharply once exceeded. The risk premium depends on the nature of the collateral asset. If the collateral is more stable assets like USDT, ETH, or WBTC, the risk premium will be very small or even zero. Conversely, high-risk meme coins will have a much higher risk premium to prevent “good assets subsidizing bad assets.”

A simple example: in V3, interest rates depend entirely on supply and demand. Borrowing USDT, even if the maximum loan-to-value (LTV) and liquidation thresholds differ, the interest rates for collateralized ETH and LINK are the same under the same supply-demand conditions. However, LINK’s volatility is higher than ETH’s. If interest rates are equal, borrowers collateralizing LINK increase utilization, which could cause the borrowing cost for ETH collateral users to rise instead of fall.

V4 addresses this flaw. Borrowers using high-risk assets as collateral must pay higher costs, and lenders can earn higher returns. Meanwhile, higher interest rates restrict borrowing demand, making the cost advantage for users borrowing with high-quality collateral more apparent.

Regarding liquidation, liquidators will only restore the health factor to the target value preset in the Spoke. The lower the health factor, the higher the liquidation bonus. This design not only provides borrowers with greater operational flexibility but also reduces the platform’s bad debt risk. Additionally, the new liquidation engine introduces a “dust prevention mechanism”: when remaining debt or collateral falls below a threshold (e.g., $1000), liquidators must liquidate the entire position to prevent small residuals from accumulating and lowering capital efficiency.

Finally, idle liquidity within the Hub can be automatically invested into governance-approved low-risk yield strategies (such as short-term government bonds, stablecoin liquidity pools, money market instruments, etc.), increasing returns for liquidity providers while also boosting DAO revenue. This may be one of the few advantages of “unified liquidity.”

Overall, the benefits of Aave V4’s unified liquidity in lending and borrowing are not particularly significant, and the so-called composability—where borrowers can manage positions across different Spokes in a unified manner—does not offer much more convenience than V3. But as the author mentioned in the title, V4 transforms Aave from a mere product into a financial infrastructure akin to a “bank.”

Setting aside complex business operations, the core function of a bank is to accept deposits, keep a portion as reserves for daily payments and transfers, and earn the interest spread through lending. Idle funds can also be allocated to various investments within the bank’s risk appetite.

St. George Palace, headquarters of St. George Bank

Founded in 1407 in Genoa, Italy, St. George Bank is generally considered the world’s earliest bank. It not only provided deposit and loan services but also handled government debt management, currency exchange, and fund transfers, meeting the commercial needs of Genoa as a major European trade hub at the time.

From launching ETHLend in 2017 to the release of Aave V4 in 2026—less than ten years—Aave has evolved into something resembling a traditional bank. Of course, Aave and a bank are quite different; this is only an analogy. Compared to P2P models, a banking system that has endured hundreds of years of black swan events is naturally a better model—just as V4 is an evolution of V3.

If you observe carefully, you’ll find that much of the so-called “innovation” in DeFi has almost become history’s dust. For example, the hot DeFi 2.0 in late 2021; instead, Aave’s simple business model, with logic matured over centuries in traditional finance, has survived and thrived. After years of exploration, many DeFi projects have likely realized this: DeFi’s ceiling is high, but it cannot skip the foundational steps that traditional finance has taken.

Aave V4 consolidates liquidity, opening up many possibilities for the future. For example, assets that remain idle for over a certain period (such as one year) could be used for higher-risk investments, like providing ETH/USDT liquidity on Uniswap, operating entirely like a commercial bank, and gradually expanding into other banking services such as credit cards (similar to Ethfi’s model of using collateralized loans for stablecoin spending), and more.

Furthermore, Aave could expand into an “investment bank.” For instance, launching an ICO platform where users depositing assets to earn interest could also lend USDT or USDC to participate in investments—without needing to withdraw and sell assets to obtain stablecoins for ICO participation. This way, the platform could collect fees from projects and earn interest simultaneously.

Although the Hub & Spoke mechanism itself does not bring major innovations to lending and borrowing, it lays the most important groundwork for the next steps.