Aave V4 vs V3: How Hub & Spoke Architecture Transforms Risk Exposure

The DeFi lending landscape is about to shift. Aave V4, currently in final security audits and set to launch in early 2026, introduces a radical architectural redesign that replaces V3’s market-per-pool model with a Hub & Spoke system. For liquidity providers, borrowers, and governance participants, this means a complete rethinking of how capital efficiency, risk isolation, and liquidation mechanics work in practice.

This article breaks down the core architectural changes and explores what they mean for your risk exposure on Aave.

Introduction: From Market-Per-Pool to Hub & Spoke

Aave V3 built its lending markets on an isolated market-per-pool model. Every new asset pair required separate liquidity bootstrapping—a process that fragmented capital across markets and made it expensive to launch specialized lending pools. If you wanted a separate market for stETH with different risk parameters than the main ETH market, V3 forced you to bootstrap entirely new liquidity pools from scratch.

V4 fundamentally inverts this design. Instead of fragmented pools, V4 introduces a Hub & Spoke architecture: a centralized Liquidity Hub that holds all protocol assets and system-wide accounting, paired with modular risk-isolated Spokes that implement specialized borrowing logic. The Hub provides shared liquidity infrastructure. The Spokes enforce distinct risk parameters without fragmenting capital.

This distinction matters enormously. V3’s distributed model prevented systemic concentration but created massive capital inefficiency. V4’s unified Hub enables capital to flow freely between Spokes—without requiring users to migrate their funds or governance to coordinate pool-by-pool upgrades.

The trade-off: greater capital efficiency now comes at the cost of introducing a new systemic risk: hub concentration. All protocol liquidity depends on a single, shared Hub.

Architecture Fundamentals: How Hub & Spoke Works

To understand how V4 isolates risk without fragmenting liquidity, you need to grasp the technical division of labor:

The Hub is a centralized liquidity repository. It maintains unified asset custody for all tokens (ERC-20 contracts), system-wide accounting, and the ability to enforce protocol-level circuit breakers. When you supply WETH to Aave V4, that WETH lives in the Hub, not in a market-specific pool. The Hub also issues Hub Assets—ERC-20 tokens that track your proportional claim to shared capital—enabling Spokes to access liquidity without requiring migrations.

Spokes are modular borrowing modules, each with independent risk parameters, oracle interactions, and emergency stops. A Spoke doesn’t hold assets; instead, it borrows from the Hub (via a credit line) and supplies to the Hub (via a debit line). This decoupling means:

New Spokes can launch without existing liquidity migration
Each Spoke can specialize for different asset classes (stablecoins, staked ETH, high-risk assets, leveraged strategies)
Hub governance sets baseline parameters; Spokes configure autonomously within those bounds
A Spoke’s default can be contained without affecting other Spokes or the Hub itself

The architectural elegance lies in optionality. In V3, adding a new market meant coordinating governance, bootstrap mining, and pool-by-pool integrations. In V4, a new Spoke can launch with access to all Hub liquidity from day one. You, as a liquidity provider, don’t need to do anything—your capital is already available to the new market.

Risk Isolation Without Liquidity Fragmentation

V3’s risk management toolkit included Isolation Mode (siloing collateral to specific markets) and Siloed Borrowing (preventing high-risk assets from being used as collateral system-wide). These tools worked but came with a cost: they created fragmentation. Isolated markets couldn’t freely borrow from other pools, leading to inefficient capital allocation.

V4 replaces this with unified Spoke-level controls. All assets remain stored in the single Liquidity Hub per network. But Spokes enforce distinct risk parameters:

A stablecoin Spoke might allow any stablecoin as collateral
An institutional-grade Spoke might accept only WETH, wstETH, and Aave’s governance token (AAVE)
A high-risk asset Spoke might specialize in liquidating emerging tokens with higher price volatility

Each Spoke can set its own collateral weights, liquidation thresholds, and borrow rates. The innovation: governance changes apply only to future positions. If you already borrowed UNI against your WETH holdings under the old risk parameters, a governance vote to increase UNI’s risk premium won’t liquidate your existing position. You’re grandfathered in. Only new borrows trigger the updated parameters.

This elegance prevents the governance death spiral seen in V3: borrowers panic-sell when risk parameters are tightened, triggering cascading liquidations. V4’s forward-only approach separates governance updates from position protection, achieving granular risk segregation without forced liquidations.

Dynamic Risk Premiums: Collateral Quality Matters

Here’s where Aave V4 moves beyond V3’s uniform rate structure with a fundamental economic redesign.

In V3, all borrowers in the same market paid the same interest rate, regardless of the collateral they pledged. If you borrowed GHO against WETH (one of the safest collaterals), you paid the same rate as someone borrowing GHO against UNI (far more volatile). This created a subsidy mechanism: users with high-quality collateral implicitly subsidized users with risky collateral. It was economically inefficient and politically contentious—why should WETH holders help UNI borrowers?

V4 introduces a three-tier risk premium system:

Asset Liquidity Premiums (0–1000% per asset, assigned at listing): WETH might have a 0% premium, LINK a 30% premium, UNI a 40% premium
User Risk Premiums (weighted average of a user’s collateral mix): Your personal premium depends on the quality of your specific collateral basket
Spoke Risk Premiums (aggregate measure across a Spoke’s users): Reflects the overall risk profile of a Spoke

The practical effect: a 5% GHO base rate might mean:

WETH collateral users pay 5% (0% premium)
LINK collateral users pay 6.5% (30% premium applied)
UNI collateral users pay 7% (40% premium applied)

This redesign kills the subsidy mechanism. Borrowing costs now reflect actual collateral risk. It also creates sustainable DAO revenue—higher premiums flow to the protocol treasury—and directly incentivizes users to keep high-quality collateral.

Liquidation Engine Redesign: Target Health Factor Over Fixed Close Factor

V3’s liquidation model was simple: when a borrower’s health factor fell below 1.0, liquidators could repay exactly 50% of the borrower’s debt and claim a fixed liquidation bonus (typically 5–10%). This fixed close factor worked as a blunt instrument. If a position was worth $100,000 and dropped to a health factor of 0.8, liquidators would repay $50,000 and claim $5,000 in collateral as their fee—whether or not that much liquidation was actually needed.

V4’s Target Health Factor system inverts this logic. Instead of specifying how much debt to repay, governance specifies a target health factor (e.g., 1.05) and a maximum liquidation bonus (e.g., 10%). Liquidators repay only enough debt to restore the borrower’s health to that target. If restoring health to 1.05 only requires repaying $30,000, that’s all the liquidator can claim. The borrower loses less collateral.

The bonus scales dynamically. As health factor falls, the bonus increases—a Dutch auction structure that incentivizes urgent liquidation when positions are truly distressed. Three governance parameters control this:

maxLiquidationBonus: Maximum bonus (e.g., 10%)
healthFactorForMaxBonus: Health factor threshold where bonus hits maximum (e.g., 0.95)
liquidationBonusFactor: Rate at which bonus scales as health falls

One more innovation: dust positions (very small outstanding balances) must be fully liquidated. If a position falls below a governance-set DUST_LIQUIDATION_THRESHOLD, liquidators can clear it entirely, preventing zombie positions that clog protocol state.

The result: Aave V4 prevents over-liquidation, protects borrowers from unnecessary collateral loss, and ensures the protocol doesn’t accumulate unprofitable dust.

Account Representation: From Rebasing aTokens to ERC-4626

V3 used rebasing aTokens to represent your deposit. When you supplied WETH, you received aWETH. Your aWETH balance automatically increased each block as accrued yield was distributed. If you started with 1.0 aWETH earning 5% APY, after a year your balance would grow to 1.05 aWETH. The increase in token balance reflected accrued interest.

This design creates problems for DeFi composability. Smart contracts that integrate with Aave must track both balance and balance history—any external protocol monitoring your balance needs special logic to detect yield accrual separately from voluntary supply increases.

V4 shifts to ERC-4626 share accounting. Your balance stays static. Instead, yield is tracked via a rising price-per-share. You supply 1.0 ETH and receive 1.0 shares. Yield doesn’t change your share count; instead, each share’s value rises. After a year at 5% APY, your 1.0 shares is now worth 1.05 ETH. Total ETH in the Hub is the same, but your proportional claim has increased.

Why does this matter?

DeFi composability: External protocols can integrate without special balance-tracking logic
Tax treatment: Many jurisdictions treat rebasing as a distribution event; ERC-4626 simplifies tax reporting
Operational simplicity: Yield is implicit in share price, not explicit in token supply

V4’s tokenization (aToken issuance) remains optional and is built on top of ERC-4626 vaults, so integrations can choose their preferred abstraction level.

Risk Exposure Changes: Hub Concentration vs. Distributed Pools

Now, the critical question: does V4 reduce or increase overall risk for participants?

The answer is complicated. V4 trades one form of risk for another.

V3’s distributed market-per-pool architecture meant risk was geographically fragmented. A crash in one isolated market couldn’t cascade to others—bad debt in the stETH market didn’t affect the WETH market. But fragmentation came at a cost. Liquidity was thin and scattered. New markets bootstrapped slowly. Capital couldn’t flow flexibly to where it was needed.

V4’s unified Hub solves the capital efficiency problem. All liquidity flows through a single pool, enabling any Spoke to access any asset instantly. But it introduces a new systemic risk: hub concentration. If the Hub’s smart contract code contains a vulnerability, or if some black-swan liquidation cascade overwhelms the Hub’s accounting, it affects all Spokes simultaneously. There’s a single point of failure.

Mitigating this risk requires careful governance:

Emergency stops per Spoke: Each Spoke has independent guards and can be frozen without affecting others
Hub-level circuit breakers: Governance can pause the entire Hub during systemic stress
Conservative governance: Hub concentration risk demands conservative baseline parameters and active parameter calibration

The trade-off is explicit: V4 accepts higher systemic risk (hub concentration) in exchange for higher capital efficiency. Governance must remain vigilant.

Governance Implications: Active Parameter Calibration Required

V3’s liquidation model was static. Close factor was fixed globally at 50%. Liquidation bonuses were fixed. There was little for governance to debate—the mechanics were locked in.

V4 demands active, ongoing governance. Because liquidation parameters are now per-Spoke and governance-set (Target Health Factor, maxLiquidationBonus, liquidationBonusFactor), the community must continuously calibrate these settings. Higher targets protect borrowers but reduce protocol revenue from liquidation fees. Lower targets maximize fees but increase over-liquidation risk.

Similarly, risk premium assignments (the 0–1000% scale for each asset) require informed governance decisions. Is LINK risky enough to justify a 30% premium? Should UNI’s premium increase to 50%? These decisions now flow directly through Aave’s governance portal.

This is a feature, not a bug. V4’s configurability enables precision—you can tune parameters for specific Spoke use cases—but it demands expertise. Governance delegates must develop deep domain knowledge about liquidation mechanics, collateral risk, and spoke-specific dynamics.

The implication: Aave V4 governance becomes more complex but more powerful. Parameter decisions that were once immutable are now levers for optimizing protocol health and revenue.

Conclusion: A Riskier but More Efficient Protocol

Aave V4 represents a maturation of DeFi lending design. By unifying liquidity in a central Hub and isolating risk in modular Spokes, the protocol achieves capital efficiency that V3 could never reach. New markets launch instantly. Governance can configure specialized pools without user migration. Liquidation mechanics are more precise and borrower-friendly.

The cost: systemic risk concentrates in the Hub. Parameter tuning replaces static mechanics. Governance participation becomes mandatory, not optional.

For liquidity providers, V4 offers higher capital efficiency and higher APY potential — though actual returns depend on Hub utilization and market conditions — but requires closer monitoring of hub-level governance decisions. For borrowers, V4 offers more precise collateral pricing and protection from over-liquidation—but introduces exposure to hub concentration risk. For governance participants, V4 demands active, informed calibration of a far richer parameter set.

Aave V4 is not a simple upgrade. It’s an architectural rethinking that trades certain risks for others. Understand the trade-offs before you participate.