Hubs & Spokes in Aave v4

Introduction

The Aave V4 architecture opens a new design space for structuring lending markets. The Hub and Spoke model moves the protocol beyond the monolithic pool design of V3, in which each market was a self-contained silo with a single risk profile. V4 allows each Spoke to define its own risk appetite, collateral policies, and liquidation rules while drawing on shared Hub liquidity. We can view it as a supranational bank allocating capital to regional facilities, each operating under its own mandate.

In this publication, we present various market configurations enabled by V4’s architecture and discuss the trade-offs among them, ranging from pooled setups that resemble V3 to segmented Spokes per collateral type to fully isolated Hubs in which collateral is never re-hypothecated.

We cover mechanisms that allow finer risk differentiation. Risk Premiums allow Spokes to price borrowing costs based on collateral quality; borrowers using riskier assets pay for the risk they introduce, rather than being subsidised by those posting ETH or BTC. Credit Lines between Hubs allow controlled liquidity flow, and when layered as a waterfall, they bootstrap new markets while capping contagion through draw caps at each level.

DISCLAIMER: This analysis does not provide risk parameterisation recommendations or evaluate which assets should be listed. The scope is limited to mapping the architectural design space and its risk implications.

Market Concepts with GHO

Gho Reserve Distribution Architecture

GHO is minted by a GHO Facilitator into a central GHO Reserve, serving as the single source of GHO supply across the protocol. From the Reserve, GHO is distributed through independent channels, each with its own governance-controlled add cap, for example:

  • Lending Credit Lines to Prime, Aave-aligned Hub, Ethena, EtherFi, and Neobank Hubs, making GHO available for borrowing within each Hub’s Spokes.
  • stataUSDC GSM and stataUSDT GSM for peg stability through direct swap mechanisms that are separate from the liquidity markets.

Each channel operates independently, so governance can scale GHO supply into high-demand Hubs while keeping tighter limits on newer environments without one affecting the capacity of another.

Gho Stewards interact exclusively via the Credit Line funded by the Gho Reserve, with the Add Cap acting as the primary constraint on GHO expansion. This mirrors the economic role of the Gho Facilitator in Aave v3, where capacity is managed at the reserve level while the Add Cap replaces the Borrow Cap as the control on total GHO exposure. Governance and Gho Stewards retain direct authority over system growth through Credit Line adjustments.

With each channel having its own cap and operating in a distinct market environment, governance can also choose precisely which collateral/risk profiles support incremental GHO supply (i.e., the asset exposure the protocol takes). This allows Aave to scale GHO where the underlying collateral and liquidation conditions are most trusted, while limiting exposure to newer or riskier setups.

Separating stablecoin liquidity across distinct capped channels provides granular control over liquidity composition, liquidation behaviour, and surplus generation, while isolating each Spoke from broader Core market risk, allowing the Core market to provide a secondary-market savings option that complements sGHO, which has a suitably lower risk profile.

Concept - Prime Hub

The Prime Hub best represents a more Senior market, characterised by isolated lending pools with tighter risk controls. Whilst a simple, isolated market structure can be created, the example above highlights how the Reinvestment feature could provide a distinct advantage relative to other markets. There are a few options for allocating the excess capital; one possibility is Simple Earn on Coinbase, or, to minimise counterparty risk, an allocation to a suitably low-risk Hub-and-Spoke, such as the Coinbase Hub-and-Spoke shown later. Creating such a market in which the supply yield always exceeds that of T-bills would be a welcome improvement relative to the current market, where supply rates are well below T-bill yields. The intention for this market is to limit collateral types to those with the lowest complexity, such as fiat-redeemable stablecoins and the most dominant BTC wrapper products.

The graphic shows Lido’s wstETH, the most dominant Liquid Staking Token (LST) with widespread adoption, as yield-generating collateral with the Max. Lido Yield Spoke is configured to support leverage yield strategies such as the stVault product. By isolating LST and Liquid Restaking Tokens (LRT) across the Hubs-and-Spoke, for example, Core and Prime, the market can price the risk associated with each collateral type. This also allows the cost of borrowing ETH to be market-determined without the need for directional/opinionated use of the Risk Premium between LRT and LSTs.

Note: Lido’s wstETH is used as an example, and other suitable LSTs may be included.

Concept - Core Hub with a Reinvestment Feature

The Core market presented above is risk-adjusted and favours yield-generating collateral types, capital efficiency, and the creation of passive yield for the user, most similar to the existing Core market on Ethereum, with specialised Spokes, Draw Caps, and Add Caps used to limit asset exposure.

Aave v4 expands the DAO’s ability to express liquidity and risk policy at the system level, rather than having outcomes emerge solely as a by-product of utilisation dynamics. Idle stablecoin liquidity in the Core Hub is no longer structurally constrained to lower productivity; it can be allocated, within governance-defined bounds, into highly liquid, low-risk instruments, thereby establishing a baseline return profile for idle capital.

This approach enhances capital efficiency without altering the protocol’s liquidity assurances. By allowing a portion of idle capital to remain productive within governance-defined constraints, v4 supports a more stable return profile than prior versions, translating design flexibility into measured economic benefits at the protocol level.

Reinvestment Feature

Stablecoins sitting idle in the Core Hub (not currently borrowed by any Spoke) are put to work. A reinvestment controller routes them into yield strategies such as sUSDS to earn the Sky Savings Rate, Coinbase’s Simple Earn, and potentially highly liquid RWAs. At scale, the reinvestment feature can behave like a portfolio to balance duration schedules and target a specific yield profile.

The Core Hub also extends credit lines to other Hubs (Aave-aligned, EtherFi, Ethena). When borrowing demand increases, reinvested capital from yield strategies is immediately withdrawn and made available. The interest rate model accounts for the amount of reinvestment, so rates adjust accordingly.

Risks and considerations

Bad debt on any Spoke is shared across all Hub suppliers. Draw Caps are the primary safeguard, bounding maximum loss per Spoke. Reinvestment is only safe with instant-redeem strategies. If too much idle capital is reinvested and multiple Spokes spike utilisation simultaneously, there may not be enough instant liquidity to serve all withdrawals and borrows at once.

Concept - Neobank Configuration (EtherFi)

This configuration places a dedicated EtherFi Hub at the centre of a Neobank use case, in which users deposit weETH as collateral and borrow stablecoins that are loaded onto an EtherFi Pay card. weETH continues to earn restaking yield while serving as collateral, thereby allowing users maintain ETH exposure and staking returns while accessing stablecoin liquidity.

Stablecoin liquidity enters the Hub through two paths. USDC and USDT are sourced from the Core Hub via a stablecoin credit line. GHO enters from the GHO Reserve through a GHO Credit Line with a defined Add Cap, acting as a backstop, ensuring the market remains funded even if the stablecoin credit line reaches its cap. Users interact with one of the EtherFi Pay Spokes, which handles collateral, borrowing, and position tracking, and draws stablecoins from the Hub via a Stablecoins Draw Cap.

Borrowers who draw GHO can route it through the Gho Stability Module to receive either USDC or USDT, enabling fiat conversion and supporting any payment card. GHO’s borrowing rate is also likely to be less volatile due to the inherently flatter utilisation curve and can be customised through governance to support specific partnership arrangements.

Whilst this Hub would likely be a net receiver of stablecoins from the Core Hub via the credit line, any idle USDC or EURC could be routed to Coinbase’s Simple Earn product, acting as a low-risk staking option. The yield could be either directed to attract new deposits, routed to Aave DAO to enhance revenue generation, or used to reduce the EtherFi Foundation’s opportunity cost of providing capital to the market.

Risks and Considerations

Collateral is concentrated in a single asset, weETH or eBTC, in a single collateral-spoke configuration. Any de-pegging or slashing event would affect all positions in the absence of a diversification buffer. Stablecoin availability depends on Core Hub credit line and GHO Reserve caps, both governance-controlled. If both reach their limits during periods of high demand, borrowing may be temporarily constrained, though this is partially mitigated by the GHO Stewards, who can mint and deposit additional GHO on short notice.

Concept - Centralised Exchange Hub (Coinbase)

Three isolated Spokes each pair a single collateral asset (cbBTC, cbETH, wETH) against USDC as the sole borrowable asset. Isolating each collateral ensures that a value impairment of one asset does not affect positions backed by other collateral types at the Spoke level. USDC liquidity enters through a Prime Hub stablecoin credit line, reflecting the lower risk profile of this Hub-and-Spoke combination. Idle USDC is routed into Coinbase Simple Earn, the native yield strategy for the Coinbase/Circle ecosystem, though governance could configure any instant-redeem alternative.

To ensure the On-chain Earn yield from Coinbase exceeds the passive Simple Earn yield, the reinvestment feature, in combination with some account mapping by Coinbase, could direct a portion of the Simple Earn yield to On-chain Earn users to amplify the capital efficiency of the system. Such a system can be implemented using the Centralised Exchanges API and would amplify the On-chain Earn user case whilst still enhancing Aave DAO’s revenue relative to the case with no reinvestment feature.
Risks and Considerations

A Coinbase operational issue could impact both cbBTC and cbETH simultaneously, even though they sit in separate Spokes. Because USDC is the only borrowable asset, all three Spokes depend on a single Prime Hub credit line.

Concept - Stablecoin Issuer Hub (Ethena)

Ethena holds stablecoin reserves as part of USDe’s backing, and during periods of low funding rates, USDe’s backing is allocated to DeFi yield-generation strategies, such as sUSDS and depositing into various lending markets. The Stablecoins Spoke allows assets to be allocated into a purpose-curved Ethena Hub as supply-only deposits. To reduce Ethena’s concentration risk, a Credit Line is provided from the Core Hub and a GHO Credit Line is extended via the Gho Reserve as a backstop, such that Ethena can supply/withdraw from the stablecoin spoke with ease.

The two isolated cbBTC and wETH spokes indicate how Ethena is currently allocating a portion of USDe’s backing to lending protocols that utilise isolated market structures. Each spoke can access liquidity for USDT, pyUSD, USDC, and GHO via Draw Caps.

The Borrow Rate, set at the Hub, can be configured with a Steward/Risk Oracle to ensure that the Borrow Rate is relative to the Reinvestment feature rate of return, thereby ensuring that any assets backing USDe generate at least a specified ratio of the Reinvestment feature’s returns.

In this example, we use the Sky Savings Rate (SSR) for reference only. The market has a borrow rate linked to sUSDS and a premium equal to the Aave Fee. This is similar to how Spark deployed USDS to the Aave Prime instance, where the Borrow Rate was a function of the SSR plus a premium because it was not the Spark Protocol itself.

Linking the Borrow Rate to the Reinvestment feature ensures that Ethena always earns at least the SSR. However, any asset with suitable liquidity could be used, such as a T+1 RWA earning over 5%, with additional considerations needed for managing the liquidity requirement. The larger GHO becomes, the easier it will be for GHO to facilitate providing the liquidity needed to ensure the market is always well capitalised. The user would need to migrate debt from another asset to GHO; however, this would be routed via the GSM to ensure a seamless user experience.

In this market configuration, we demonstrate that the yield from the Reinvestment feature enhances returns on the assets backing USDe relative to actively allocating directly to the strategy or to a lending protocol, which has no floor yield (i.e., no reinvestment feature). Aave’s revenue would be derived from the Reserve Factor on the debt, which would be expressed as a premium relative to the selected investment strategy if all the yield is directed to Ethena.

Risks and Considerations

Backing assets serve dual roles as USDe redemption capital and Aave liquidity. During a contraction, mass redemptions may prompt withdrawals from the Hub as borrowing demand shifts. Bad debt on either Spoke is shared across all Hub suppliers, including Ethena’s own deposits.

Concept - Aave-aligned Hub

This configuration establishes a dedicated Aave Aligned Hub containing two isolated Spokes, designed for borrowers who utilise protocol-aligned assets as collateral to access stablecoin liquidity. The Aave Spoke supports directly aligned collateral (stkAAVE, AAVE), while the Chianlink Spoke supports indirectly aligned collateral (LINK). Collateral across both Spokes is not rehypothecated and remains continuously available for withdrawal and liquidation, ensuring solvency restoration depends solely on spoke-level liquidation mechanics and enabling controlled tuning of risk and execution outcomes.

Both Spokes support borrowing USDC and USDT from the Core Hub, alongside GHO introduced through a dedicated Gho Credit Line, each governed by an independent Add Cap. Whilst USDC and USDT provide diversity, the Gho Borrow Rate can be configured at a lower rate than that of other hubs to recognise stkAAVE and AAVE holders as part of a broader tokenomics benefits package.

Given the borrower-aligned design of this Hub, a borrower-UX–oriented liquidation configuration, featuring smaller liquidations and lower liquidation bonuses, should be considered across both Spokes. This approach minimises borrower disruption, reduces collateral loss, and maintains predictable liquidation behaviour, aligning with the Hub’s objective of supporting protocol-aligned participants while preserving strong capital backing and controlled risk.

While the collateral sets are subject to asset volatility, both Spokes intentionally bias slightly toward borrower UX relative to what volatility alone might suggest. This is enabled by strict isolation from the Core Hub, conservative add caps, and spoke-level risk containment, which together allow for a more gradual liquidation profile without materially increasing bad debt risk. In stressed scenarios, faster execution parameters remain available as a governance-controlled adjustment, allowing risk management to be tightened if market conditions warrant.

Risk and considerations

The Aave Spoke’s collateral base is structurally correlated, as stkAAVE and AAVE are both closely linked to the Aave ecosystem and may be affected simultaneously by protocol-level events. The LINK Spoke introduces independent risk vectors LINK’s value is driven by Chainlink network adoption and broader market conditions rather than Aave-specific dynamics. The two-spoke architecture ensures risk events in one Spoke do not propagate to the other. Risks across both Spokes are mitigated through conservative caps, spoke-level isolation, and governance-controlled parameters.

Conclusion

This post outlines concepts of possible Hub-and-Spoke configurations enabled by V4. We covered:

  • Prime Hub: A lower risk, Senior market focusing on isolated use cases and offering controlled exposure to the most established multi-cycle assets, with a focus on fiat-backed stablecoins
  • Core Hub: Pools blue-chip collateral across multiple spokes, providing a risk-adjusted yield with elevated capital efficiency relative to v3 via the Reinvestment features.
  • Etherfi, Coinbase Hubs: Opinionated market participant hubs focused on specific use cases, optimised for each respective business’s considerations
  • Ethena Hub: A low-risk, isolated market structure could be positioned to generate a yield on assets that provide collateral backing for a stablecoin
  • Aave-Aligned Hub: Isolated Spokes for protocol-aligned assets like stkAAVE, AAVE, wstLINK & LINK with preferential cost of capital alignment via a Gho Credit Line

Each design entails different trade-offs among capital efficiency, risk isolation, and liquidity depth.

GHO is integral to every configuration as the backbone stablecoin across V4 markets. Each Hub draws GHO from the Reserve through independently capped Credit Lines, so governance controls exactly how much GHO enters each market and against what collateral. As these Hubs scale, GHO scales with them.

It’s important to note that none of these are final recommendations. Risk parameterisation and asset listing decisions are controlled by AAVE token holders and elected Stewards, typically specialists, to facilitate operational implementation.

Disclaimer

TokenLogic did not receive any payment for this post.

Copyright

Copyright and related rights waived via CC0.

8 Likes

nice

1 Like

Thanks for the analysis, this is great work.

On the Prime Hub, for the launch I would keep it focused on the simple “no-rehypothecation” setup where volatile assets can be used as collateral and stablecoins can be borrowed, and I would exclude (at least initially) the Max Lido Yield / leverage-yield spoke. We can always introduce that later once the v4 is more established and governance has more operational comfort.

Also, because collateral is not rehypothecated, I don’t see a strong advantage in splitting Prime into very specific asset-class spokes (i.e. BTC/USD mode). Any incremental LTV improvement versus a grouped spoke would likely be very small at cost of a UX fragmentation. On the other hand, I really like the reinvestment feature in the Coinbase Simple Earn hub; it’s a very elegant structure and feels like a repeatable pattern that could also be combined with similar hubs designed for prime institutions.

On the Core Hub, I strongly agree with single spokes for LSTs / LRTs.
It’s the most effective way to keep risk policy clean, increase LTV over time, and give the DAO more granular control through draw caps. For the same reason, I would remove the mixed LST/LRT spoke at launch. I understand the point that ~25% of LSTs/LRTs loopers deposits comes from multi-collateral positions (weETH + wstETH for the most), but adding a blended spoke early introduces again UX complexity and professional loopers will always be able to achieve the same outcome by using two spokes, and they may even get better effective LTV than a mixed configuration. We can always re-evaluate adding a more extended spoke later if the demand is strong enough.

Finally, I would not place Maple and Ethena inside Core. I would put them in dedicated hubs and let Core support them through credit lines. This keeps Core’s risk simpler and gives those hubs a clearer growth instrument without forcing Core to internalize their full risk surface.

The Neobank configuration resonates a lot, I think there are many opportunities to grow this segment across multiple players, and EtherFi is a great example. Overall, Coinbase and Neobank feel like strong near-term opportunities, while the Ethena and Aave-aligned setups are interesting but probably better suited for a later phase. In particular on the Aave-aligned design, I’m not fully convinced we need a dedicated Hub; since it’s not a distinct liquidity domain, I believe a Spoke configuration would fit better than a separate Hub.

Great work again.

3 Likes

Great work, @TokenLogic team, as always.

These are very interesting, out-of-the-box ideas.

What stands out clearly from this proposal is the structure: Prime as a low-risk hub, Core as risk-adjusted, and three more opinionated markets. Among the three, Ethena follows a similar approach to what was previously presented with risk providers. Ether.fi and Coinbase serve as good examples of how hubs can be tailored to specific use cases. The Coinbase-focused hub, for example, could eventually align well with a Base deployment down the line.

For the genesis launch on mainnet, I would recommend keeping things simple with Prime, Core, and a single risk-return hub. Use-case–tailored hubs and spokes can then be workshopped and developed based on these foundational concepts.

Overall, this configuration demonstrates how GHO growth could be driven by seeding credit lines across all hubs and spokes, while especially ensuring a strong liquidity backstop.

At this point, the community has at least three different providers presenting approaches to the Hub and Spokes architecture, which gives a solid foundation to build on. We will consolidate all proposals into a single post to improve readability before the DAO proceeds with a decision on a follow-up proposal once there is enough discussion.

2 Likes