- Two tracked Aave Umbrella stablecoin stakes on Ethereum V3 fell by $12.4 million between September 17 and October 17, 2026.
- Their extra reward rates rose over the same period, with USDC moving from 1.66% to 1.82% and USDT from 1.46% to 1.63%.
- Balances are dollar-valued snapshots, so the decline cannot by itself show whether funds were withdrawn or lost.
Two of Aave’s tracked Ethereum V3 stablecoin backstops shrank by $12.4 million in the month to October 17, 2026, even as the additional rewards paid on them increased. The combined USDC and USDT positions stood at $116.09 million at the later observation, down 9.67% from $128.52 million at the earlier one. The data comes from DefiLlama’s position histories and reflects dollar values rather than raw deposit counts.
What Happened
The positions sit within Umbrella, Aave’s system that uses staked assets to absorb lending deficits. Stakers continue to earn the underlying lending yield and also receive an extra reward for accepting the risk that their funds could be used to cover losses. DefiLlama identifies the two tracked positions as stkwaEthUSDC.v1 and stkwaEthUSDT.v1, and their token identities match Aave’s current address book.
The USDC stake declined from $58,739,617 on September 17 to $53,437,993 on October 17, a reduction of $5.30 million. The USDT stake fell from $69,783,937 to $62,655,694, a reduction of $7.13 million. Both comparisons use snapshots taken at 23:01 UTC on September 17 and 15:01 UTC on October 17, 2026, so the interval is slightly less than 30 days.
Over the same period, the extra reward component rose by 0.16 percentage points for USDC and 0.17 points for USDT. The base lending rate also moved, with USDC’s rising from about 3.60% to 3.77% and USDT’s from about 3.02% to 3.12%. The analysis focuses only on the extra component, since that portion compensates stakers for the risk of slashing rather than reflecting lending returns.
Why It Matters
Multiplying each tracked balance by its extra reward rate gives an indicative annualized figure of roughly $1.994 million. That amount is nearly identical at both observations, at $1,993,923.12 in September and $1,993,859.28 in October. In other words, a higher percentage rate was paired with a smaller pool of capital, while the implied aggregate payout barely changed.
For investors, this combination suggests that the incentive offered per dollar staked became more generous even as less capital was committed to the protection layer. The analysis treats these numbers as indicative run rates derived from snapshots. Actual accrual depends on how rates, stake values, reward settings and participation evolve over time.
The balance decline also should not be read in isolation. The figures measure dollar values, so movements include changes in asset prices and in the wrapped-aToken exchange rate, not only changes in deposits. A lower balance therefore cannot identify whether stakers withdrew funds or whether losses occurred.
Market Context
Umbrella’s reward controller uses an emission curve tied to the ratio of staked capital to a configured target. Aggregate emissions peak at the target and decline between the target and 120% of target, eventually flattening at or above that 120% boundary. Above that threshold, fewer stakers share the same reward stream, which mechanically raises the rate per unit of capital. The documentation makes this mechanism clear, but the historical data alone does not confirm whether it explains the movements in these two pools. That would require matched contract settings and balances at both dates.
Reward allowances are another piece of context. A September 25 proposal from TokenLogic sought to renew the allowances used to pay Umbrella rewards. It left emission rates, target liquidity, cooldown parameters, deficit offsets and the December 2, 2026 distribution end unchanged. The renewed allowances cover rewards already accrued but not yet claimed, including amounts owed to former stakers, along with remaining emissions through that end date. Because of this, the allowance figures should not be treated as a fresh annual incentive budget.
Protection rules also shape how stakers should interpret the numbers. Each reserve is backed by its own stake: USDC stake covers USDC deficits, and USDT stake serves the USDT reserve. Slashing is triggered only when a realized deficit exceeds the applicable deficit offset, which LlamaRisk describes as a DAO-funded first-loss layer. Aave documents a 20-day cooldown followed by a two-day withdrawal window. Rewards continue to accrue during cooldown, and the stake remains exposed to slashing while the holder waits. Aave’s separate V4 coverage plans involve different protection modules and are not part of these V3 observations.
What to Watch Next
- Stake balances in the coming snapshots: Whether the USDC and USDT positions continue to decline, and whether any change reflects deposits, withdrawals or valuation effects.
- Reward rate adjustments: Whether the extra rates keep rising and how they relate to the emission curve and target levels.
- Governance and allowance activity: Any further proposals affecting reward allowances, emission parameters, cooldown rules or the December 2, 2026 distribution end.
- Deficit events: Any realized reserve deficit that would test the deficit offset and could trigger slashing.
Source: original report