Aave is not freezing Ethena USDe. On 2 September 2026 at 20:09 UTC, LlamaRisk posted a Risk Steward interest-rate note that raises the USDe base variable rate by 100 basis points on five V3 deployments and lifts Slope1 by 20 basis points on fifteen other stablecoin reserves. The destination named in that post is still a 5.25% USDe base, not a 3% finish line.
This is a rate path with receipts. The same note says the USDe base already moved from 0.00% to 1.00% on 28 August and from 1.00% to 2.00% on 1 September. The new spec puts Core and Avalanche at a 3.00% base, with Slope1 cut from 4.00% to 3.00% on those two markets so the borrow APR does not overshoot the 5.25% to 5.5% band TokenLogic mapped in August.
Ether sat at $2,507.38, up 4.98% over 24 hours, on the up&down CoinGecko snapshot fetched at 2026-09-03T23:55:17+00:00. That tape does not decide this story. The fee and the curve do.
What happened
LlamaRisk, writing as a DAO-funded risk provider, said it will “move forward and implement these updates via the Risk Steward process.” That is the same pipeline we covered when USDe borrow caps were raised, not frozen. Caps answer how much USDe debt Aave will house. This note answers how that debt is priced.
On USDe, the recommended base variable rate is 3.00% on Aave V3 Core, Plasma, Monad, Mantle, and Avalanche, up from 2.00%. Core USDe utilization in the post is 54.4% against a 90% optimal point. New Core borrow APR is 4.81%, up 40 basis points, because the Slope1 cut offsets part of the base increase. Plasma, Monad, and Mantle see a full 100 basis-point APR step. Avalanche, already running hotter at 68.8% utilization, only moves 14 basis points to a 5.58% APR.
On other stables, Slope1 goes from 4.10% to 4.30% on fifteen reserves that LlamaRisk says hold about $5.83 billion of debt. Core USDC sits at 93.0% utilization versus a 94% optimal point. Core USDT is 92.6%. Several satellite markets (Avalanche USDC and USDt, Monad USDC, MegaETH USDm, Plasma USDT0) are already at or through their optimal points. Base rate, optimal utilization, and Slope2 stay put on those fifteen. The squeeze penalty above the kink therefore shifts up by the same 20 basis points.
The migration table is the part I care about. From the last snapshot before the first USDe base step through the latest snapshot in the note, Core USDe debt fell $107.0 million across 86 accounts. $56.8 million of that (53.1%) was re-borrowed as USDC or USDT by the same accounts. Only about 4% of the Core reduction came from accounts that left the market entirely. Plasma saw $14.3 million of USDe debt come off, with roughly 8% of that from accounts that closed out. This is rotation inside Aave, not an emptying of the pool.
Context
TokenLogic’s 24 August thread, August 2026 stablecoin interest rate adjustments, is the parent document. It prices Ethena yield-bearing stables against native staking rates, stages Slope1 higher where utilization can pay for it, and treats the USDe move as a sequence, not one vote. LlamaRisk is now on step three of that USDe base ladder (0% to 1% to 2% to a recommended 3%), still well short of the 5.25% base named in both posts.
Show me the fee, not the roadmap. DefiLlama’s Aave fee summary, fetched 2026-09-04 at 00:20 UTC from the Llama fees endpoint, printed $1,155,414 in 24-hour fees, down 4.14% on the day, and $8,020,670 over seven days. Ethereum chain fees on the same desk pull were $9,337,789 in 24 hours, down 12.93% day over day. Those prints can move with ether’s tape. They do not, by themselves, tell you whether a 20 basis-point Slope1 bump on $5.83 billion of near-kink stablecoin debt is “tight” or “cosmetic.”
Protocol TVL on DefiLlama’s Aave series last point dated 2026-09-03T23:12:59+00:00 showed about $18.44 billion of total liquidity. The chain breakdown in that same pull put Ethereum TVL near $15.64 billion with $10.50 billion borrowed. That is a large book. It is also why a steward can change a rate without waiting for a full governance circus, and why the published path matters more than a single day’s APR delta.
Ethena USDe circulating supply was about $4.26 billion (pegged USD) on the DefiLlama stablecoins feed fetched 2026-09-04 at 00:20 UTC, with a printed price of 0.99998. Aave is one venue among several for that float. Pricing Core USDe at a 3% base while Core USDC already sits at 93% utilization is how you stop USDe from remaining the cheap borrow of first resort after two prior steps already pushed some of that debt into USDC and USDT.
If you want the longer Ethereum history or a plain definition of a stablecoin, those hubs are on the desk. For how we timestamp a tape versus a protocol print, use how to read a crypto market snapshot. More of this beat lives under Ethereum news. TVL is a size number. It is not a rate.
Our read
I read this as a scheduled repricing, not a freeze and not a finished 5.25% destination. Updating our view from the 1 September cap note: the cap raise said Aave still wants USDe size. This IRM note says Aave will not keep that size at a 2% base while USDC and USDT sit on the kink. Those two claims fit. One is inventory. The other is price.
The Core math is the tell. Utilization on USDe is only 54.4%, so a 100 basis-point base increase would have overshot the TokenLogic band unless Slope1 came down. LlamaRisk cut Slope1 on Core and Avalanche for that reason. On the fifteen other stables, they left the kink and Slope2 alone. That preserves the squeeze incentive. It is not a stealth freeze of withdrawals. Optimal utilization is unchanged, so the share of supply sitting as exit liquidity at the kink is unchanged.
Falsifiable claim: by 2026-09-09 23:59 UTC, a LlamaRisk or Risk Steward primary (forum spec, payload note, or Aave UI parameter print) shows the Aave V3 Core USDe base variable rate at 3.00%, matching the 2 September recommended spec. If Core USDe base is still 2.00% at that deadline with no steward delay note, the “implement via Risk Steward” sentence in the 2 September post is wrong. If a TokenLogic or LlamaRisk primary withdraws the 5.25% USDe base-rate path before that deadline, the destination half of this read is wrong.
What to watch next
First, the Core USDe base print. 3.00% is the next published step. Until it hits a payload or the UI, this remains a recommendation with a steward promise, not a live curve.
Second, Core USDC and USDT utilization after the 20 basis-point Slope1 bump. They were 93.0% and 92.6% in the note, against a 94% optimal point. If utilization falls hard while USDe base is still climbing, the rotation story in the $107 million table is still in force. If they stay pinned to the kink, Slope1 is doing income work, not emptying the book.
Third, whether LlamaRisk keeps marching the USDe base toward 5.25% in 100 basis-point clicks. Two steps are already done. A third is queued. A skip, a pause note, or a jump would change the path.
Fourth, the Aave fee print on DefiLlama. A $1.16 million 24-hour fee day (2026-09-04 00:20 UTC fetch) can absorb a 20 basis-point Slope1 nudge. A collapse in that print alongside a utilization crash would say demand, not the curve, is the story.