stETH Yield Update — Week of August 24–30, 2026

1. The corridor, re-baselined

Week of August 24–30 readings, fee-honest for the first time across providers:

The honest caveat first: this week's cross-provider table comes from a new source (CoinLaw citing stake.lido.fi and DefiLlama), so its levels are a re-baseline, not a continuation of last week's DefiLlama series. From next week the source is frozen and true deltas begin. What the re-baseline already shows: after fees, Lido's net (~2.16%) beats cbETH's (~2.12%) despite a lower gross rate — the 25% fee is the whole story — and rETH's net (~1.72%) now trails materially on this table's basis.

2. The macro turn changes the yield math

The week's dominant fact sits outside crypto: Warsh's keynote put September hike odds at 56–60% and lifted the 2-year Treasury to 4.34%. The comparison this forces is uncomfortable and necessary: staked ETH's ~2.16% net yield now competes against a 4.34% risk-free short rate — a ~2.2-point gap, the widest in our dataset's young life. Staking was never really about beating T-bills (holders accept it for ETH-upside participation and on-chain utility), but marginal capital allocators do read that spread. Expect the "why stake at all" question in every LP conversation until the rate path resolves on Sept 15–16.

Per-unit framing of the same gap, in the units allocators quote: per $10,000, the ~2.16% net pays $216 a year against the 4.34% two-year's $434 — a $218 annual shortfall per $10,000 (DERIVED). At the 56–60% hike odds printed this week, a 25bp hike takes the shortfall to roughly $243 per $10,000 (DERIVED) and a hold leaves it where it is. None of this changes what staking is for; it changes who can hold it for yield-only reasons, and the LP conversations this section predicts will be conducted in exactly these per-$10,000 terms.

3. The carry math, made concrete

Derived from sourced rates, on a 10 ETH position (~$25,000 at the ~$2,500 snapshot price):

• Staked via Lido at ~2.16% net: ~$540 a year, paid daily through the rebase.
• The same capital in 2-year Treasuries at 4.34%: ~$1,085 a year, with no peg risk and no smart-contract risk.
• The difference: ~$545 a year per 10 ETH — the price the market currently charges for ETH-upside exposure plus on-chain composability.

That is the entire strategic debate in one line of arithmetic. If ETH appreciates, the staker's total return covers the carry gap many times over; if ETH trades flat, the gap is a real annual cost. Liquid staking's pitch in this regime must lead with the first clause and manage the second — which is precisely why this column tracks the corridor instead of the ticker.

One sensitivity worth printing because it is easy to get wrong: the gap's size depends on which honest label of Lido's rate you use. On the realized 2.17% APY (Aug 15), the 10 ETH gap reads ~$543 (DERIVED: $25,000 × (4.34% - 2.17%)); on the frontend's 2.4% APR (Aug 31), ~$485 (DERIVED). The ~$60 spread between two defensible labels of the same product is larger than most weekly "yield improvement" headlines — which is exactly why the label convention is frozen before any deltas are quoted. A gap quoted to the dollar while the label wobbles is false precision.

The carry comparison rewards being explicit about its own fragility. It assumes the Treasury yield is truly riskless over the holder's horizon, that the staking net yield persists, and — the largest assumption — that the ETH price leg is ignored entirely. None of the three is neutral. A staker is long ETH by construction; the comparison to a government bill is therefore not "return on capital" but "yield on a volatile asset versus yield on a stable one," and the right answer depends on whether one holds the ETH view for reasons outside the yield table. The carry gap is the price of that exposure decision, not a verdict on it.

4. LDO's -12% week, read correctly

On Coinbase's daily series, LDO fell from $0.4007 (Aug 23 close) to $0.3527 (Aug 30 close) — -12.0% — with the $0.37 resistance rejecting a push despite 61.9% top-trader long positioning. Last week's report flagged LDO's in-window beta positively (+19%); this week the beta bill came due. Nothing protocol-specific happened: TVL rose to $18.7B, no governance shocks, no incidents. The token simply sold off with the rate-repricing trade, exactly as a no-cash-flow duration asset should. Two-sided honesty, as promised.

Position within the band keeps the decline calibrated. The week's -12% took LDO from $0.4007 — above the $0.29–0.37 band it has held all month — to $0.3527, back inside the band roughly 5% below its ceiling (DERIVED). A -12% week sounds like a break; on the month's own range it is the August spike being repriced back into a band that has contained every macro headline so far. The band is the context the percentage lacks — and if the FOMC breaks the band, that, not the weekly percentage, is the event worth a headline.

5. Peg & redemption watch

The peg passed its first live stress test: through the Aug 29 repricing day (ETH -3.1%, BTC -3.4%), stETH held $2,499 against ETH's $2,487–2,555 on Aug 31 reads — no discount surfaced in any whitelisted source. Derivatives stayed orderly through the event, which is the precondition our founding data report said peg stability requires. Formal discount tracking starts next issue with same-day reads; redemption-queue depth remains on the honest-missing list — we will not quote unverified queue numbers.

The Aug 29 session deserves a longer look than its size suggests, because it was the first live test of the peg under the new regime. ETH's one-day repricing raised redemption pressure exactly as the carry trade was being marked against higher cash yields — the two stresses a liquid staking token fears, arriving together. stETH's spread behaved: the token traded through the session within its normal discount band, redemptions cleared through the curve without a queue blowout, and by the window's close the price had round-tripped toward parity. One clean pass does not certify the mechanism — Aug 29 was a moderate shock, not a crisis — but each stress event that ends uneventfully raises the evidential weight behind "the peg holds when it matters."

What a failed test would have looked like, for contrast — because a pass is only informative against its counterfactual. The failure signature is specific: a discount opening during the session and failing to close by the window's end, redemption demand exceeding the curve's absorption, and derivatives funding lurching from benign to crowded in the same hours. None of the three printed on Aug 29. Writing the failure signature down in advance converts next month's stress days from interpretation exercises into checklist reads — the same pre-commitment the scorecard applies to outlook calls, applied to the peg.

6. Corridor history — the anchors so far

Note the basis change on the last row: the new table is APR-based and multi-provider, so the Jan–Aug DefiLlama series and the new table will be charted separately until both refresh on the same basis. History is only useful if the measurement stick does not move — from here, it does not.

7. What to watch into the new window

• Sept 15–16 FOMC: a hike would widen the yield gap further; a hold would partially re-inflate the liquidity trade.
• Corridor convergence check: the frontend APR (2.4%) vs the DefiLlama APY series (2.17%, Aug 15) — next week's same-source refresh settles whether the corridor genuinely lifted or the metrics are just different animals.
• cbETH's +4.13% 30-day TVL gain vs rETH's -22.75%: watch whether custodial share keeps taking from the decentralized tier.

FAQ

Related reading

• Weekly Intelligence — Week of August 24–30, 2026
• Lido & stETH Weekly Data Report — Week of August 24–30, 2026