[Risk Stewards] August 2026 - Stablecoin Interest Rate Adjustments

[TokenLogic]

---

title: [Risk Stewards] August 2026 - Stablecoin Interest Rate Adjustments
author: @TokenLogic
created: 2026-08-24

---

Overview

TokenLogic proposes updating Aave stablecoin interest-rate curves. The proposal raises Slope1 where recent utilization or persistent supply and demand conditions support a higher target rate, moderates Slope2 on X Layer to reduce borrow-rate volatility during periods of growing demand, aligns the borrowing cost of Ethena yield-bearing stablecoins with their native staking rates, and adopts a temporary UOptimal change to house the demand transfer from Ethena interest rate changes. The changes seek to preserve competitive borrowing conditions while ensuring that stablecoin debt is priced consistently with its use across Aave.

General slope increase

Stablecoin rates need to reflect both sides of each market. Borrow demand can support a higher rate when users continue to take on debt at the existing target, while declining or flat supply can justify a higher return when demand remains stable or grows. We applied these principles over a persistent observation period so that a temporary movement on either side of a reserve does not determine the recommendation. The following 22 reserves are proposed for a 50 bps increase in Slope1.

GHO remains under separate rate management. Frozen reserves remain unchanged.

At current borrow balances and utilization, these changes add an estimated $2.55 million of annual DAO revenue if demand remains unchanged. This is a static estimate, and the staged execution outlined later in the analysis allows the DAO to evaluate the response before every subsequent increase.

Unlike the other markets, we recommend limiting the change on Base USDC to 25 bps. Its Slope1 already stands at 4.50% and would move to 4.75%, leaving the market at a target rate above the other Aave deployments. The market remains meaningfully lower cost than alternatives on the same chain and can command a premium over comparable chains.

One of the reserves in scope, USDT on BNB Chain, operate meaningfully below their optimal utilization. On BNB Chain, debt has continued to grow, demand remains solid, and the constraint sits on the supply side, so a higher supplier return is the appropriate lever to attract the deposits this demand supports.

Economic context

Ethereum, Plasma, and Base are three of Aave’s most significant deployments; as such, they require additional consideration when changes are applied. For this reason, we assess the broader rate environment on each chain to ensure that Aave remains competitive for sizeable borrows and leveraged strategies. The comparison uses Aave borrowing costs, borrowing incentives where applicable, the largest suitable lending-market comparator, and the staking rates available from the yield-bearing stablecoins listed on each chain. Every series will be presented as APR.

Aave Borrow Rate vs On-Chain Comparison - Ethereum

Aave Borrow Rate vs On-Chain Comparison - Plasma

Aave Borrow Rate vs On-Chain Comparison - Base

On Ethereum, Aave is currently 25 to 60 bps cheaper than alternative venues for any volatile borrow demand, and comparable to Spark Lending. Following the change, the Aave market will command a considerable premium over its two main competitors. This could prove counterproductive by shifting demand; the borrow rate change will be implemented gradually and closely monitored.

The spread between the relevant yield-bearing stablecoin rate and the Aave borrowing cost leaves roughly 5% of room for maximum leverage. However, these leveraged-strategy yields are expected to rise above 5% as the underlying yield environment improves, further strengthening Aave’s market position and pushing the leveraged yield to 7-8%. Plasma has no lending market comparator of sufficient size at this time, so its chart compares Aave borrowing costs against the available sUSDe and syrupUSDT rates. That spread leaves roughly 6% at maximum leverage after accounting for the live USDT0 borrowing incentive. On Base, Aave remains roughly 25 basis points cheaper than the next comparable venue, while the yield-bearing spread leaves roughly 8% at maximum leverage.

Expected market response

The projection below estimates how the Ethereum Core stablecoin reserves absorb the combined effect of the staged Slope1 increases and the USDe realignment. It rests on a set of assumptions that are listed as follows:

• On the supply side, USDC deposits are assumed to grow by $14M per week, in line with the market’s observed behavior over the past month. USDT deposits are assumed to grow by $18M per week once reserves reach optimal utilization, mirroring the growth observed for USDC, provided sufficient demand reaches the market, and the broader environment becomes supportive of USDT borrowing.
• Yield-bearing collateral rates are held at current levels, 4.2% for syrupUSDT and 4.8% for syrupUSDC, while sUSDe is carried at the 5.3% rate derived later in this analysis.
• Borrowers are assumed to migrate between lending venues only when the rate advantage exceeds a tolerance, defined as rate sensitivity, quantified from historical behavior, as observed migration patterns show that borrowers tend to stay put when the difference is minimal and tend to assign premiums and discounts to certain venues.
• The borrow demand released by the USDe borrow rate change is assumed to be redistributed between other Aave markets where liquidity is available, filling each reserve up to its optimal utilization.
• Within Aave, the two reserves are observed to trade at a sustained spread, reflecting increased demand for USDT borrows and the USDT weak market peg.
• Deposits, in turn, are assumed to migrate from Spark Savings whenever the Aave supply rate exceeds their hurdle rate by a sufficient margin. Suppliers tend to react more slowly than borrowers, but Spark Savings remains the most likely venue from which additional supply is attracted in an isolated market environment, and the sensitivity is kept consistent with the rest of the framework.

Ahead of this projection, we conducted a statistical analysis on historical rate and parameter changes to estimate how the supply side of these markets reacts to them. The exercise did not yield a clear correlation, as deposit reactions could not be separated from ordinary market variability over the time horizon this rate-change program operates on. As such, this analysis covers the supply side only in a limited fashion, through the relationship between Aave supply rates and the Spark Savings products. The numbers presented below should be read in that context: the supply side is likely to react more positively than simulated, but the quantification of its difference is unmodeled.

Under these assumptions, the expected reaction at each step of the sequence is the following.

Ethereum Core USDT

Ethereum Core USDC

Most of the adjustment is expected to concentrate in the first step. The borrow demand arriving from the USDe rate change is met by an estimated $366M of deposits migrating from Spark Savings to the USDT market, whose hurdle rate is reached at the UOptimal level, and the reserve absorbs the inflow without a lasting rate spike. At the same step, roughly $10.1M of syrupUSDT-collateralized loops are expected to close, as the higher rate leaves them below the return a levered position requires.

Across the remainder of the sequence, an estimated $44.8M of volatile-collateral USDT debt is expected to migrate to Spark before capacity there is filled. USDC crosses no external threshold at any step; its movement comes from within Aave, as borrowers switching over from USDT fill the headroom its supply growth creates over time and hold the reserve near its optimal utilization through most of the sequence. The spread between the two reserves widens with each step until it reaches the USDC rate premium as stated in the initial assumptions.

We expect the market to land in a significantly stronger position than it started. The combined book ends roughly $408M larger than today, split between roughly $356M on USDT and $52M on USDC once the switching between the two reserves is accounted for, with the USDe demand absorbed and repriced at market rates. Realized borrow rates are expected to settle slightly below the published targets, at approximately 4.27% on USDT and 4.47% on USDC, because the supply the program attracts eases utilization as it arrives. Total outflows across the sequence, near $55M between migration and closures, remain small next to the retained and newly attracted demand.

The projection is an estimate, and the staged execution is intended so that each step can be analyzed for negative reaction before the next one is taken.

Slope2 adjustment on X Layer

Leveraged positions depend on borrowing costs remaining manageable as demand grows quickly. The current configuration on X Layer can cause high volatility in borrowing rates during these periods, forcing positions to reduce debt abruptly even when the underlying strategy remains viable. LlamaRisk covered the effect in its recent analysis of the Monad stablecoin reserves, recommending an increase in optimal utilization from 90% to 92% and a reduction of Slope2 from 40% to 20%. We support these values and recommend extending the Slope2 reduction to X Layer, which runs the same configuration. The changes lower the maximum borrow rates on these reserves from 44.00–45.00% to 24.00–25.00%. While the parameters remain above the configurations used on the more established deployments, we recommend maintaining a buffer that reflects the novelty and size of these instances.

Yield-bearing alignment

We also recommend aligning the borrowing rates of USDe with their staking rates so that users cannot borrow these assets at rates lower than their native protocols allow. A lower borrowing floor can otherwise support positions that collect rewards while paying an inadequately priced debt cost on Aave. The proposal sets the USDe base rate at 5.25%, and the standard Slope1 at 0.25%.

The higher borrowing floor is expected to close a material share of current USDe debt. Most of this debt sits in positions that recycle borrowed USDe back into sUSDe, a structure that carries overlapping incentive costs for the wider Ethena ecosystem. As these positions unwind, we expect the Ethena Staking Rate to rise by roughly 50 bps, leaving both the Ethena product and the Aave positions built on it healthier. A higher staking rate also improves how other Aave markets respond, driving growth expected to offset the closed positions.

USDe borrower response and expected migration

Ethereum Core currently has $459.8M of USDe debt. Volatile-collateral borrowers account for 50.5% of the instance, while borrowers using yield-bearing stable collateral account for the remaining portion. Other positions contribute minimally. We separated wallets in the market using a 90-day activity cut to distinguish dormant debt from active positions. Roughly $1.25M of the volatile-collateral segment lies in the first group.

Plasma has a different composition, with 99.9% of its $154.2M USDe debt sitting in the yield-bearing stable segment.

Monad and Mantle add $36.0M and $2.3M of USDe debt and hold no yield-bearing stable collateral at all, with Monad at 82.1% plain-stable collateral and Mantle entirely volatile.

The proposed rate will affect the two economically distinct uses of USDe differently.

Current USDe debt by borrower sensitivity segment across Ethereum Core, Plasma, Monad, and Mantle.

The current curve allows loop-funded deposits to earn protocol yield while the associated borrowing remains cheap, creating overlapping incentive costs without adding external demand for USDe. The proposed 5.25% APR base rate links the borrow cost to Ethena’s current staking rate after accounting for Aave’s reserve factor.

Loop-funded deposits dilute the yield available to other Ethena holders. Based on our estimate, this change in the interest rate curve should improve the annual yield accrued by the Ethena system by roughly 50 bps.

If loop-funded supply unwinds, we expect the Ethena staking rate to reach roughly 5.3% APY. However, the uplift depends on the extent of the unwind and the resulting contraction in sUSDe supply.

Expected change in the borrower book

Following this change, active loop positions would be the first to respond, migrating shortly after the rate increase. Given the increase in Ethena Staking rate, transferring the open positions to stablecoin debt will remain highly profitable. Such migration will increase utilization and demand in the destination markets.

These borrowers against volatile collateral are likely to be the last to move and the least likely to leave. And when they do, they are also the most likely to refinance into other stablecoin debt, which would also add utilization and demand to those destinations.

In the nearer term, the remaining USDe book would therefore be expected to include the dormant wallets and a material share of this sticky volatile-collateral debt. Over a longer horizon we expect this debt to migrate as well, though the pace is uncertain and the outcome not assured.

The two migration paths have distinct economics. As such, the proposed USDe curve could remove the current leveraged loops even if volatile collateralized borrows adjust more slowly or remain.

Stablecoin UOptimal Temporary Increase

As the USDe realignment is expected to move a meaningful portion of its debt toward cheaper stablecoin markets, the immediate constraint sits in the liquidity available in the USDC and USDT reserves on Ethereum Core and Plasma. To accommodate the migration, we recommend a temporary 2% increase in the optimal utilization of these reserves. At current deposit levels, the change frees roughly $116M of additional borrowing capacity, allowing the market to absorb demand from USDe before the supply side has time to react to the supply rate increase.

The purpose of the adjustment is to avoid spikes in the borrow rate of the market stablecoins while the books absorb the arriving demand. It is intended only as a temporary measure: once the market attracts sufficient liquidity following the Slope1 increase, the optimal utilization is expected to be walked back to its current levels together with the rest of the parameter changes.

Slope1 staged implementation

We recommend implementing the Slope1 increases in 10 bps steps through the Risk Stewards, with all in-scope reserves moving together. This process reflects reserve dynamics. Persistent utilization near the kink establishes that borrowers value the market liquidity above its current price, but does not establish the ceiling of this rate. A 10 bps overshoot is visible in the data and simple to unwind, while a single full-sized move would reveal a misjudgment only after the demand it priced out had left.

Borrowed amount is the primary indicator between steps because debt directly measures whether demand agrees to the new price. While debt holds or grows, the premise behind the increase remains intact, and the slope1 increases progress. A contraction, however, one that exceeds the reserve’s normal variability and is not observed in comparable markets we have left untouched, indicates the ceiling has been found: past that point, further increases would cost more in departing debt than they earn in borrowing rate. TokenLogic will stop further increases on the reserve if the above is observed.

Deposit growth is the process objective. Our primary objective with the rate increase is to attract additional supply to clear the demand bottleneck. However, it takes time for depositors to establish that the new, higher supply rate is stable and worth migrating to. This behaviour is visible historically, as deposit growth has lagged the rate change. We therefore expect supply to react more slowly than debt. When, following an increase, no deposit growth nor debt shrinkage has happened, we will pause the schedule and allow time for that response to materialize before increasing further.

Utilization acts as the balance between the two sides of the market. If deposits do arrive faster than borrowing grows into them, the utilization is pushed below the range the reserve normally occupies. Each reserve therefore carries a floor drawn from its own utilization history. A reserve below its floor has, for the moment, more capacity than demand, and further rate increases would serve no purpose until borrowing absorbs the new supply. The schedule will then suspend for that reserve and resume once utilization recovers within the target.

USDC on Ink and USDT on BNB Chain instead require a different consideration. As both were selected below their optimal utilization, with solid demand constrained by shallow supply, for these two reserves, deposit growth is itself the measure of success, and our target is to primarily maximize that.

While we target each 10bps change to be executed weekly, if we observe the market reacting positively and responding faster than expected, we can increase the speed at which these increases are executed, up to the Risk Stewards’ limit of 72 hours.

We expect individual weeks to show little change during the process. A 10 bps step sits inside the daily variability of these markets. The cumulative pattern should be used to determine progress. Every week, TokenLogic will assess internally each reserve against its pre-existing trends and comparable unchanged markets, so that a market-wide movement is not attributed to the program.

When the sequence reaches the full 50 bps Slope1 increase, or 25 bps for the Base USDC market, the market situation will be reassessed to determine whether another proposal is necessary.

Execution

The base-rate changes for USDe execution will be conducted through Risk Stewards in order to expedite the start of the rate adjustment. While these moves exceed what a single Risk Steward action can carry, and moving from zero to its aligned level would require five to six increments with a cooldown between each, the gradual implementation allows us to observe the first effects of the change in the days following the proposal. In order to align the migration of USDe positions with the increase of UOptimal, we recommend backloading the adjustment to Slope 1. The complete execution is expected to stretch over the following two weeks.

For the Slope2 and optimal utilization changes, we recommend executing them after publication of this proposal through the Risk Stewards process, in sync with the execution of the first change to the USDe base rate.

Specifications

The following specifications cover the first parameter changes recommended in this analysis. Specifications for the next changes will be posted as a response to this post.

Slope1

USDe Alignment

Slope2 Reduction

Stablecoin UOptimal Adjustment

Forward look

Further analysis will cover the stablecoin reserves not addressed in this proposal. The remaining work is concentrated in smaller markets and in reserves where current conditions may support rate reductions. Those changes require separate treatment because liquidity, lifecycle status, and the response to prior rate changes differ materially across the remaining books.

Disclaimer

TokenLogic is an active service provider to the Aave DAO, the beneficiary of stream 100086 and the KPI as outlined in this publication. The scope of this engagement is available via this forum proposal.

TokenLogic supports and maintains an independent delegate voting platform within the Aave community.

TokenLogic and associated entities have no undisclosed material conflicts of interest at the time of submission.

Next Steps

1. Expand the specification of LlamaRisk’s proposal to the full Slope 2 and UOptimal set presented above.
2. Gradually implement the Slope 1 and Base changes via Risk Stewards, followed by close monitoring.

Copyright

Copyright and related rights waived via CC0.

[CryptoScam]

How big was the bribe coming from Ethena grifters to rug USDE borrow?

Surprisingly the revenue increase is mentionned for the bluechip adjustments but the sharp reduction in usde revenue is nowhere to be seen…

Is that governance or bribe show?

[9PashaE6ashu]

I’m opposed to the USDe section of this proposal as written. The Slope1, Slope2 and UOptimal changes look reasonable on their own merits; the USDe realignment rests on a number that is never derived and on two projections that cannot both arrive at the scale the model assumes.

1. The 50 bps uplift is asserted repeatedly and derived nowhere - and the model cites a derivation that doesn’t exist

The figure appears in the Yield-bearing alignment section as an expectation that the Ethena staking rate rises by roughly 50 bps, and again in the borrower-response section as an improvement to the yield accrued by the Ethena system, attributed to your own estimate with no inputs shown. The ~5.3% APY endpoint follows immediately, caveated by the extent of the unwind and the resulting contraction in sUSDe supply.

Meanwhile, the Expected market response assumptions carry sUSDe at 5.3% and describe that rate as derived later in the analysis. It isn’t. The later passage is the estimate plus the caveat above. The model’s key input cites a derivation that appears nowhere in the post.

2. The inputs the projection implies don’t match observable data

The only way to reconcile a 5.3% endpoint with a 50 bps uplift is an unstated current staking rate of ~4.8%. That input appears nowhere in the post either, and it doesn’t match what’s observable: sUSDe has averaged around 4% over recent months, and the incentive layer is visibly underfunded - the Merkl distributions promised on USDe are not being paid in full, and against a ~4% target (parity with sUSDe), realized USDe yield is running near 2.5%, of which incentives contribute only ~1.36%. If the ecosystem cannot fund parity for unstaked USDe today, there is no visible funding source for a further uplift on sUSDe.

Start from the observable ~4% and one of your two numbers has to give: either the uplift really is ~50 bps and the endpoint is ~4.5%, not 5.3% - or the endpoint really is ~5.3% and the uplift is ~130 bps, which would require loop-funded deposits to make up a far larger share of total sUSDe supply than anything evidenced here.

Before any execution, please publish: the current staking rate used as the base, the observation window (spot or trailing average, and over what period), the assumed loop-funded share of sUSDe supply, and the reserve-factor adjustment behind 5.25%.

3. The unwind projection and the migration projection require the same positions to do mutually exclusive things

A loop position has two exits, and everything downstream depends on which one it takes:

• Unwind: unstake sUSDe (with a cooldown), receive USDe, repay. sUSDe supply contracts → the uplift can exist. But the debt leaves Aave entirely; nothing migrates.
• Migrate: borrow USDT/USDC, buy USDe on the market, repay. Demand arrives in the stablecoin reserves - but the sUSDe stack is untouched, supply doesn’t contract, and the uplift never happens.

The proposal books both outcomes from the same positions, each at full scale. The uplift - and the healthier-Ethena framing - needs the unwind. The demand-redistribution assumption, the temporary +2% UOptimal (~$116M of capacity explicitly reserved for migrating USDe demand), and the projected ~$408M book expansion all need the migration.

The migration case is also justified circularly: transferring positions to stablecoin debt is said to remain clearly profitable because the staking rate rises. But by construction the uplift arrives only after the unwind — a borrower deciding whether to migrate faces today’s ~4% sUSDe, not the projected 4.5–5.3%. Against your own ~4.27% projected USDT rate, that spread is roughly zero or negative net of incentives; the rational response is to close, not to migrate. The migration forecast stands on a number that the migration itself prevents from materializing.

The asymmetry is visible in the model’s own inputs: syrupUSDT and syrupUSDC are held at current levels (4.2% and 4.8%), while sUSDe alone is carried at your post-intervention forecast of 5.3%. The one input marked to your own projection rather than to market is the one that justifies the program.

4. Even on its own terms, this is a premium over native yield, not alignment

The base rate is quoted as 5.25% APR; Aave rates compound, so borrowers pay ≈5.39% APY. At every utilization level, borrowing USDe would cost more than the asset’s native yield.

5. This doesn’t reprice the USDe book — it evicts it, on a liquidation-like clock

Nobody holds USDe debt at a ~5.39% floor while functionally equivalent stablecoin debt sits near 4.3%. Outside marginal niche demand, the book doesn’t reprice - it exits. That includes the 50.5% of the $459.8M Ethereum Core book held against volatile collateral, which the proposal itself describes as the stickiest segment and still expects to leave over time. Reserve revenue on USDe accordingly trends to zero - not because the rate is unpaid at the margin, but because no one remains to pay it - while recapture into USDC/USDT rests on the migration case shown above to be self-undermining.

And the exit window is punitive. Five to six Risk Steward increments over ~2 weeks put the full floor in place inside a fortnight, while sUSDe’s unstaking cooldown means borrowers who need USDe to repay cannot exit through redemption on that schedule - they are pushed into secondary markets. Compressed, involuntary exits from a ~$614M book across Ethereum Core and Plasma, into thin secondary liquidity, at whatever price is available, are economically closer to a soft liquidation than to a repricing - except the losses are imposed by a parameter change rather than by market risk, and the pressure lands on USDe’s secondary price at the same time.

The track also carries none of the discipline this same proposal applies to Slope1. There, you publish explicit criteria: stop on debt contraction beyond normal variability, pause when neither deposits grow nor debt shrinks, suspend below a utilization floor. The USDe change runs to a fixed target at the maximum speed Risk Stewards permit, with no published pause or rollback conditions. A 525 bps parameter move gets less safeguarding than a 50 bps one.

6. If the problem is the loop eModes, fix the eModes

The behaviour being targeted runs through specific eModes that allow sUSDe collateral against USDe debt at high LTV. That is a parameter problem in a handful of configurations, with direct levers: cut the eMode’s max LTV to cap the achievable leverage multiple, cap USDe borrowable within it, or retire it. These hit the loops precisely and leave ordinary USDe borrowing intact. And because the binding lever sits on the collateral side, an LTV cut on sUSDe constrains the loop no matter which asset funds it - something a borrow-rate change on USDe alone cannot claim, since the same loop can simply refinance into USDT debt.

What I’d support

1. Split the USDe change out so the rest of the proposal can proceed.
2. Publish the derivation behind the 50 bps / 5.3% figures - base rate used, observation window, assumed loop share of sUSDe supply, reserve-factor adjustment - and state explicitly which exit path, unwind or migration, each projection assumes.
3. Address the loops through eMode parameters first, then measure what demand remains.
4. If a base rate is still warranted after that, set it materially lower - enough to break loop economics without exceeding sUSDe’s actual yield - and run it on the same staged, criteria-driven schedule with published pause conditions that the Slope1 section already commits to.

If the missing derivation is published and holds up, parts of this objection fall away. As written, the section forces out a functioning ~$652M book off a number that appears nowhere in the analysis that claims to contain it, in service of two projections that cannot both arrive at the scale the model assumes.

[willray]

I completely agree with you. Raising the base borrowing rate on USDe to 5.25% will mean no one will borrow. This would essentially wipe USDe off AAVE. There’s also the 102M PT-SUSDE-OCT loop leverage on PLASMA, which is impossible to exit; this would be a hard liquidation for them.

These looper cannot migrate to USDT0 because it’s unprofitable; they can only exit at a discount on the secondary market, which will kill plasma.

the tvl on plasma will reduce 50%

[danger]

increasing slope1 on usdc, usdt to make supply aprs more competitive makes sense.

Increasing usde base borrow to 5.25% is a huge change that I don’t quite understand. This would deflate susde loops, and shrink aave tvl, while destroying usde as a borrow asset – even if that is a goal it seems like it could be done more gradually.

given how drastic the usde apr change is - why not do it more gradually? raising in two or three steps while seeing market reaction would allow data on user behavior to inform the rate change better on this asset which currently has a lot of borrow on Aave.

[9PashaE6ashu]

Thanks for no response and silent rate increase accoarding to the plan. Absolutely disgusting.

[dsadsadasdasd]

this is a good review

[TokenLogic]

The first round of adjustments was executed on August 27 through the Risk Steward process, as covered by LlamaRisk in Risk Stewards: Stablecoin IRM Changes on Aave V3 / 2026.08.27.

With the market accelerating and utilization growing across the reserves in scope, the Slope1 increases will proceed at a faster cadence, in line with the staged implementation outlined above. Accordingly, this round moves Slope1 by 20 bps, doubling the step size of the first round. The specifications below cover the second round of adjustments.

USDe adjustment concerns

Several comments above focus on the USDe realignment, so we address the main points here. The repricing deliberately trades subsidized borrow volume for market health. We expect additional stablecoin demand through migration, along with a healthier USDe market more prone to growth, to offset the revenue that the repriced borrows no longer generate.

On the 5.3% reference rate: Ethena’s rewards are not confined to staked USDe. Almost all USDe is held through a qualifying channel that earns rewards, roughly $4.1B, of which $1.3B is staked in sUSDe, $1.2B is supplied on Aave, and most of the remainder sits on partner venues running Ethena funded earn programs, from exchange hold-to-earn products to retail platforms. Only the residual balance earns no rewards. Ethena funds these rewards with yield from its backing portfolio, currently around 5%. Borrowing breaks the one-to-one link between rewarded balances and backing. When borrowed USDe is staked, the same backing funds rewards twice, once on the supplier side and once on the staked tokens; when it is sold, it ends up redeemed, and rewards keep accruing on deposits whose backing has left. Either way, Ethena currently pays rewards on the full $0.64B of borrowed USDe, even though these balances exceed the supply that actually exists. The only offset is the share of borrow interest that suppliers receive. At the pre-change borrow rates and the current 4.8% staking rate, that loss runs at about $19M per year. For each reserve, the calculation is:

The repricing recovers that $19M and redistributes it across the $3.9B of supply that keeps earning, adding roughly 50 bps and taking the expected staking rate from 4.8% today to 5.3%.

This is also why a USDe borrow rate below the staking rate represents a net loss for Ethena stakers. Every unit borrowed below that rate captures a spread funded by the rewards that would otherwise reach existing holders. The lower the borrow rate, the larger that transfer, and the more the market’s growth depends on it persisting. Pricing the borrow cost in line with the staking rate removes the subsidy and leaves both the reserve and the underlying product healthier.

For existing borrowers, a debt swap is the efficient way to adjust without closing the position. Migrating debt from USDe into USDC or USDT keeps the collateral in place and moves only the liability. The same option is available on Plasma, where positions collateralized by PT-sUSDe cannot exit by selling the collateral before maturity. A debt migration from USDe to USDT0 leaves the collateral untouched, while Plasma USDT0 is part of both the Slope 1 set and the temporary uOptimal increase, which attracts new supply and frees capacity for migrated debt. The base adjustment itself proceeds in 100 bps steps, leaving time to execute the migration. Across the program, the temporary uOptimal increase frees roughly $116M of additional borrowing capacity on the reserves expected to receive the migrated debt.

Slope 1

USDe Alignment

The Aave v4 reserves in scope will follow once the Risk Stewards are deployed on v4.

Specifications for the following rounds will be posted as the sequence progresses.

Copyright

Copyright and related rights waived via CC0.

[9PashaE6ashu]

1. Whose problem is being solved, and what does Aave get?

The entire justification in your reply is Ethena-side welfare: a $19M annual loss to Ethena’s reward pool, recovered and redistributed to Ethena holders. Aave DAO’s side of the ledger — a repriced $652M book, foregone reserve revenue, forced migrations — is covered by one unquantified sentence about migration demand and future growth.

So, directly: was this repricing coordinated with Ethena, and what is Ethena committing to Aave users in return? That question is sharpened by the fact that Ethena has already cut the rewards it promised through Merkl — which brings us to the numbers.

2. The $19M cannot be reproduced at the payout rates Ethena actually pays

Your formula assumes rewards flow on borrowed balances at the full reward rate. They don’t. The Merkl campaign on Ethereum currently pays roughly 1.7% (recently raised from about 1.3%), against promised rates approximately 2x higher. The Aave supply leg — the very leg that creates the double-count in your model — is being paid at a fraction of the rate your calculation appears to assume.

I cannot reproduce $19M from the published formula at $0.64B of borrowed USDe using observable channel payout rates. The figure only approaches $19M if the full staking rate — or worse, the projected post-intervention 5.3% — is applied to balances that are actually being paid ~1.7%. Recomputed at actual rates, the recoverable leakage falls to a small fraction of $19M, and with it the ~50 bps uplift and the 5.3% endpoint that anchor the entire realignment.

3. The loop demand you are pricing out is a symptom of the broken reward program

Consider why borrowers rotated into the sUSDe/USDe eModes in the first place. Holders of unstaked USDe were promised a reward rate near parity with staking; the payouts then collapsed to the 1.3–1.7% range, costing them months of expected yield, with the additional risk of rising USDT/USDC borrow costs on top for anyone levered. The rational response was to move into the one channel that pays reliably — native staking, levered where eModes allow it. The proposal now treats that adaptation as toxic demand to be priced out at 5.25%, while the unreliability that produced it goes unmentioned and unaddressed.

4. The USDe track still has no published stop conditions

The Slope1 track has explicit published criteria: stop on abnormal debt contraction, pause absent deposit growth, suspend below a utilization floor. The USDe base-rate track, a 5.25 bps move, still has none.

5. By your own principle, alignment is reached this round — every increment after it is pure eviction

Run the numbers on the second round itself. With this 100 bps step, the USDe borrow rate lands around 4.9% (≈3.9% today plus the increment) — already above the ~4.74% the staking rate actually pays, before even accounting for compounding. Carry on every remaining loop flips negative this week, at step two of five or six. Additional changes to the curve are already required.

6. The migration you recommend has no supported execution path

The reply presents a debt swap as the efficient adjustment: keep the collateral, move only the liability. As a transaction description, that is correct. As a product, it does not exist. For a position inside the sUSDe/USDe eMode, migrating the liability is not one action but an atomic sequence: source USDe via flash loan and market swap, repay the USDe debt in full, switch the eMode category — permitted only once outstanding debt is compatible with the target category — and re-borrow USDT or USDC against the same collateral. Neither the official Aave interface nor DefiSaver supports this flow today.

In practice, the recommended exit is available only to borrowers able to write and deploy their own flash-loan migrator. Everyone else is left with precisely the paths this proposal claims to spare them: unwinding through the sUSDe cooldown, or exiting at market. On Plasma it is stricter still — by your own description, PT-sUSDe collateral cannot be sold before maturity, which makes the debt migration not the efficient exit but the only one, for a $154M book, with no supported tooling to execute it. And this compounds the timing problem: carry turns negative now, while the exit you recommend cannot yet be executed by an ordinary user.

If the expectation is that each borrower engineers this individually, the proposal should state that explicitly. Otherwise: where is the tooling? The Aave DAO funds interface development as requested by Aave Labs. A repricing program whose safety case rests on orderly migration should not run ahead of the migration path itself.

Conclusion:

I initially assumed there was at least some reasonable explanation, but now I see an obvious slap in the face to the Aave DAO and Aave suppliers—who are essentially being left to deal with their problems on their own—all while this is being funded by Ethena behind the scenes. GJ.

[TokenLogic]

The third round of the USDe interest rate adjustments has been executed through the Risk Steward process.

Following the successful migration response observed across the USDe loan book, the accelerated adjustment path has been paused on some instances. Accordingly, Slope 1 has also been reduced by 100 bps.
The specifications below cover the third round of adjustments.

Specifications for the following rounds will be posted as the sequence progresses.

Copyright

Copyright and related rights waived via CC0.

[TokenLogic]

Following the executed parameter rounds, we are sharing a comprehensive review of the market response against the projections published in this proposal.

Since execution began, the Risk Stewards have moved Slope1 on the in-scope reserves by a cumulative 30 bps, raised the temporary uOptimal to 94% on the Ethereum Core reserves, and carried the USDe base to 4.00% in four steps, with the USDe Slope1 reduction under way. The chart below shows daily supply and borrows on the three Ethereum Core reserves with each change marked.

What has followed the projection

Borrow demand held through the increases. Ethereum Core USDT borrows grew $73M and USDC borrows grew $81M between August 27 and September 6, while borrow rates moved to 4.24% and 4.27%, with both reserves below their optimal utilization. The projection expected the books to keep growing through the sequence, and they have.

USDC is absorbing migrated demand ahead of the projection: its book grew $81M against the $39M we anticipated at this stage.

The USDe realignment is proceeding as projected. Ethereum Core USDe borrows declined from $444M at the first parameter change to $274M on September 6, a 38% reduction, with a visible step at each base increase.

Retention has exceeded the projection. The model carried $35.5M of cumulative USDT debt migration to Spark by this stage of the sequence; SparkLend’s Ethereum USDT borrows instead declined slightly over the window. No measurable outflow to comparable venues has occurred, and Aave’s borrow cost remains competitive against the on-chain alternatives through the executed steps.

What lags the projection

The supply side has not responded yet. The projection assumed a Spark Savings inflow arriving as the reserves reached optimal utilization; Ethereum Core USDT deposits are broadly flat since execution began, and utilization has risen to 92.7%. The temporary uOptimal increase to 94% is providing the intended headroom while the supply response develops, and the remaining Slope1 step raises the supplier return further.

Of the USDe debt released since the first parameter change, roughly has migrated as demand on the other stablecoins. Wallets reducing their USDe debt cut $185M over the window; $94M of it reappeared as USDT and USDC debt on the same wallets, and the remaining $92M was repaid. Across the two receiving reserves, $356M of new borrowing arrived against $207M of repayments, with addresses new to these reserves being the largest source of the new demand. The migration path remains open, and we expect to continue.

Source of new USDT and USDC borrowing	Wallets	Amount
Wallets migrating debt from USDe	37	$108.1M
Existing borrowers increasing debt	1,434	$53.2M
Borrowers new to the two reserves	614	$194.7M
New borrowing, total	2,085	$356.0M
Repayments and closures	865	−$207.4M
Interest accrued over the window		$5.2M
Book growth since the first change		$153.8M

Next steps

The sequence continues, with the market assessed between steps as before. The next round of adjustments:

Stablecoin Slope 1 Adjustment

Chain	Market	Reserve	Parameter	Current Value	Recommended Value
Ethereum	Core	USDC	Slope 1	4.30%	4.40%
Ethereum	Core	USDT	Slope 1	4.30%	4.40%
Plasma	-	USDT0	Slope 1	4.30%	4.40%
Base	-	USDC	Slope 1	4.60%	4.70%
Arbitrum	-	USDT	Slope 1	4.30%	4.40%
Avalanche	-	USDC	Slope 1	4.30%	4.40%
Avalanche	-	USDt	Slope 1	4.30%	4.40%
Monad	-	USDC	Slope 1	4.30%	4.40%
Monad	-	USDT0	Slope 1	4.30%	4.40%
MegaETH	-	USDm	Slope 1	4.30%	4.40%
BNB Chain	-	USDC	Slope 1	4.30%	4.40%
BNB Chain	-	USDT	Slope 1	4.30%	4.40%
Gnosis	-	USDC.e	Slope 1	4.30%	4.40%
Optimism	-	USDT	Slope 1	4.30%	4.40%
Linea	-	USDC	Slope 1	4.30%	4.40%
Linea	-	USDT	Slope 1	4.30%	4.40%

USDe Base Alignment

Chain	Market	Reserve	Parameter	Current Value	Recommended Value
Avalanche	-	USDe	Base	4.00%	5.00%
Ethereum	Core	USDe	Base	4.00%	5.00%
Mantle	-	USDe	Base	4.00%	5.00%
Monad	-	USDe	Base	4.00%	5.00%
Plasma	-	USDe	Base	4.00%	5.00%

USDe Slope 1 Alignment

Chain	Market	Reserve	Parameter	Current Value	Recommended Value
Avalanche	-	USDe	Slope 1	3.00%	2.00%
Ethereum	Core	USDe	Slope 1	3.00%	2.00%
Mantle	-	USDe	Slope 1	4.00%	3.00%
Monad	-	USDe	Slope 1	4.00%	3.00%
Plasma	-	USDe	Slope 1	4.00%	3.00%

Copyright

Copyright and related rights waived via CC0.

[TokenLogic]

Since the first parameter change on August 28, the USDe base rate has moved to 5.00%, and Slope 1 has been reduced to 2.00% on Ethereum Core and Avalanche and to 3.00% on Mantle, Monad, and Plasma. The Ethereum Core book has responded as projected.

As set out in the proposal, the 5.25% base rate linked the borrow cost to Ethena’s staking rate at the time, after accounting for Aave’s reserve factor. The staking rate has since risen to 4.75%, in line with the uplift expected as loop-funded positions unwound, and the same methodology now reflects a base rate of roughly 6.25%. Accordingly, we are revising the target for the next change; because it follows the staking rate, we will reassess it each following week.

We therefore recommend continuing the sequence at the same cadence, with 100 bps base increases and 100 bps Slope 1 reductions per action, to a 6.25% base and a 0.25% Slope 1 on all five instances. The next round:

USDe base

Chain	Market	Reserve	Parameter	Current Value	Recommended Value
Avalanche	-	USDe	Base	5.00%	6.00%
Ethereum	Core	USDe	Base	5.00%	6.00%
Mantle	-	USDe	Base	5.00%	6.00%
Monad	-	USDe	Base	5.00%	6.00%
Plasma	-	USDe	Base	5.00%	6.00%

USDe Slope 1

Chain	Market	Reserve	Parameter	Current Value	Recommended Value
Avalanche	-	USDe	Slope 1	2.00%	1.00%
Ethereum	Core	USDe	Slope 1	2.00%	1.00%
Mantle	-	USDe	Slope 1	3.00%	2.00%
Monad	-	USDe	Slope 1	3.00%	2.00%
Plasma	-	USDe	Slope 1	3.00%	2.00%

Copyright

Copyright and related rights waived via CC0.

[9PashaE6ashu]

AHJAHHAHAHAHAHAHHA

[CryptoScam]

Looks like some people enjoy giving their ass to Ethena…
Gotta explain me how a 6.46% borrow rate is aligned to a 4.5% sUSDe APR…