Overview
During the recent compression of the broader crypto market landscape in monetary terms, the observed lending activity has exhibited signs of relative compression, which is caused not only by the reduction of notional prices across a broad range of assets, but also compression of the on-chain yields on stablecoins, LSTs and LRTs, additionally, as 2025 has marked substantial increase in Aave deployments, a significant share of capital for no-longer incentivized markets has migrated to either more lucrative, or stable instances. Broadly, these factors have had a substantial effect on the composition and utilization levels of a set of instances, with a number of assets’ caps substantially underutilized. Given these conditions, prudent risk management requires adjustments to the underutilized instances as liquidity conditions tend to follow the overall utilizations. Hence, in order to materially constrain Aave’s exposure to tail risks associated with both assets and instances, we hereby propose a broad-scale reduction of the supply and borrow caps.
Methodology
While each market typically presents special considerations in terms of user behavior patterns, asset allocation, and risk, we have decided to take a unified approach to these recommendations and have identified markets that have exhibited long-lasting reduced utilization or have been trending downward for extended periods. While the approach is generalized, we aim to provide additional analysis covering instances and specific assets in order to balance risk constraints with potential market utility.
Primary Markets
As part of the cap reductions, we have identified the largest markets as Core, which include Ethereum Core, Ethereum Prime, Base, Avalanche, and Plasma. Fundamentally, the mentioned instances saw reduced utilization levels primarily due to yield compression and/or a lack of alternative yield-generating opportunities for the assets.
Ethereum Core
Supply-side utilization compression on the Ethereum Core instance has been driven by a broad decline in demand across several secondary assets, most notably ETHx, UNI, ezETH, as well as eUSDe and sDAI. As shown in the chart, the ezETH and ETHx markets have experienced sustained decreases in utilization over recent months, reflecting the unwind of leveraged positioning and a reduced appetite to deploy these assets on mainnet amid ongoing yield compression across LSTs and LRTs. Additionally, we infer that the drop in supply cap utilization is driven by increased competition for WETH borrowing, as lower-yielding LRTs are unable to maintain profitability as the spread between the appreciation rate and borrowing costs compresses.
For assets like UNI, crvUSD, and eUSDe, the decline in supply cap utilization is more variable. Specifically, sDAI formally migrated to stUSDS, while UNI saw a genuine decline in demand as collateral, and eUSDe is no longer producing yield due to the end of the Ethereal pre-deposit campaign.
On the borrow side, utilization remains structurally subdued, with only brief episodic spikes—such as the October tBTC increase—indicating temporary positioning rather than persistent demand. Importantly, wstETH has exhibited a substantial decline, reducing overall protocol revenue. As wstETH borrowing was primarily fueled by persistent restaking yields on a set of LRTs, the compression of the rate spread proved the strategy less efficient than a number of alternatives, decreasing the posted wstETH borrowing demand. A similar shift in BTC-derivatives can be observed: fBTC, WBTC, and LBTC have been exhibiting sub 20% utilization as staking yield for yield-bearing BTC derivatives declined. Overall, Ethereum Core now exhibits materially underutilized caps across multiple assets, suggesting a relative compression in borrowing demand and migration to alternative borrowing opportunities.
Ethereum Prime
The Ethereum Prime instance, which has historically been designed for WETH-related lending activity, including leveraged staking and restaking strategies, has lost a substantial share of both supply and borrowed assets as users migrated to the Ethereum Core instance.
Specifically, the majority of the assets, with the exception of GHO, have been showing decreased utilizations in the 0–20% range. With the migration of the LRTs and wstETH to the Ethereum Core instance, the demand to participate in WETH leveraged loops has decreased. This has had a network effect on the broader utilization across the instance, resulting in the utilization levels observed below.
Base
A supply-side decline in demand within the Base instances has been observed in ezETH, which has been used primarily in leveraged strategies against wstETH. Demand for wstETH declined as isolated restaking yields compressed, making the leveraged restaking loop unattractive. The decrease in wstETH borrowing observed has coincided with a reduction in ezETH supply, further supporting the previously outlined motivation. Additionally, we would like to note that while tBTC supply and borrow utilization have increased in the past five months, the demand is still relatively far from the caps, prompting a moderate reduction.
Avalanche
Stark examples of underutilized assets on the Avalanche instance are sUSDe and USDe, which were listed relatively recently but have yet to gain traction. Given the high utility of these assets, as evidenced by their presence across other Aave instances, we recommend a moderate decrease in the supply and borrow caps to balance risk with deployment utility.
Plasma
The Plasma instance is primarily exhibiting healthy utilization levels, with only WETH borrowing showing a consistent downtrend due to the lack of LRT-driven borrowing demand. This motivation is further exemplified by the underutilization of weETH and wrsETH caps, which are currently in the 5-20% range, prompting an adjustment of the supply caps.
Secondary Markets
Secondary instances have seen the sharpest utilization drawdowns, largely reflecting incentive expiries, fragmented liquidity, and the migration of activity toward deeper or more profitable Aave deployments. Given that persistently low utilization often implies a permanent migration of capital, we recommend more aggressive cap reductions in these markets to reduce tail risk while retaining minimal operational headroom where appropriate.
Linea
Lending activity on Linea has exhibited a broad instance-level decline in recent months. While this instance ranked as one of the largest during late summer and early fall of 2025, the aggressive incentivization of activity naturally resulted in a large-scale migration of mercenary capital; as a result, aggregate supply has compressed from $3 billion to under $200 million, while borrowing decreased from $1.5 billion to $50 million. Current cap utilizations on both supply and borrow sides are primarily below 15% across a wide variety of assets. As the incentives are not expected to be resumed in the near future, we recommend drastically reducing the supply and borrow caps across low-demand assets listed on Linea.
Optimism
The primary reason for the decline in supply and borrow utilization of USDC.e is the presence of the canonical USDC on the instance, prompting a transition to the newer version of the asset.
Sonic
Within the Sonic instance, WETH borrowing utilization is currently below 1% due to low demand for the asset. Generally, WETH borrowing within the instance presents minimal utility as it currently has no LSTs or LRTs.
zkSync, Metis, and Soneium
Due to the recent proposal to formally deprecate the Aave zkSync, Metis, and Soneium instances, we are recommending adjusting the supply caps to the current levels and reducing the borrow caps to 1 across all assets. The recommendation is made as a precautionary measure to reduce tail risk before the proposal is officially enacted.
Recommendation
Market-wide notional compression and declining on-chain yields across stablecoins, LSTs, and LRTs have driven a sustained drop in lending demand across several Aave instances. At the same time, the expansion in deployments and the expiry of incentives have shifted capital toward more lucrative or more stable venues, leaving many secondary markets structurally underutilized.
In this environment, oversized supply caps add limited user utility but meaningfully increase latent tail exposure, especially in thinly utilized markets where liquidity typically tracks utilization and can disappear quickly under stress (e.g., depegs, oracle shocks, and collateral demand gaps). Accordingly, we recommend broad reductions to supply caps (and selective borrow caps) across markets with persistent or trending-down utilization, right-sizing capacity to current activity while preserving reasonable headroom. For instances proposed for deprecation, we also recommend setting borrow caps effectively to disabled levels as a precautionary tail-risk constraint ahead of formal enactment.
Specification
In this post, we outline a targeted subset of the proposed cap adjustments, prioritizing the most time-sensitive reductions where utilization is persistently low or precautionary constraints are warranted ahead of instance deprecation. We will follow up with additional cap changes in a subsequent reply to this forum topic.
| Instance | Asset | Current Supply Cap | Recommended Supply Cap | Current Borrow Cap | Recommended Borrow Cap |
|---|---|---|---|---|---|
| Plasma | weETH | 240,000 | 80,000 | - | - |
| Plasma | wrsETH | 20,000 | 5,000 | - | - |
| Plasma | WETH | 160,000 | 30,000 | 80,000 | 25,000 |
| Linea | WETH | 350,000 | 40,000 | 315,000 | 36,000 |
| Linea | ezETH | 76,800 | 35,000 | - | - |
| Linea | wrsETH | 30,000 | 5,000 | - | - |
| Linea | wstETH | 19,200 | 7,000 | 4,800 | 1,000 |
| Linea | USDC | 600,000,000 | 25,000,000 | 560,000,000 | 23,000,000 |
| Linea | USDT | 512,000,000 | 25,000,000 | 460,000,000 | 23,000,000 |
| Linea | mUSD | 20,000,000 | 5,000,000 | 18,000,000 | 4,000,000 |
| Linea | WBTC | 160 | 100 | 8 | 1 |
| zkSync | wstETH | 2,000 | 500 | 200 | 1 |
| zkSync | WETH | 3,000 | 600 | 2,700 | 1 |
| zkSync | ZK | 125,000,000 | 45,000,000 | 1 | - |
| zkSync | USDC | 3,000,000 | 800,000 | 2,700,000 | 1 |
| zkSync | USDT | 1,000,000 | 150,000 | 900,000 | 1 |
| zkSync | sUSDe | 200,000 | 110 | - | - |
| zkSync | wrsETH | 700 | 1 | - | - |
| Metis | m.USDC | 10,000,000 | 400,000 | 9,000,000 | 1 |
| Metis | WETH | 2,300 | 150 | 720 | 1 |
| Metis | METIS | 600,000 | 80,000 | 1 | - |
| Metis | m.USDT | 4,500,000 | 250,000 | 4,000,000 | 1 |
| Metis | m.DAI | 200,000 | 25,000 | 180,000 | 1 |
| Soneium | USDC.e | 8,000,000 | 1,300,000 | 7,200,000 | 1 |
| Soneium | WETH | 800 | 600 | 720 | 1 |
| Soneium | USDT | 5,000,000 | 30,000 | 4,500,000 | 1 |
Next Step
We will move forward and implement these updates via the Risk Steward process.
Disclosure
Chaos Labs has not been compensated by any third party for publishing this AGRS recommendation.
Copyright
Copyright and related rights waived via CC0.
