Weekly Intelligence — Week of October 5–11, 2026

1. Staking health

The question this week is whether a 6.9% price drawdown damages staking, and the answer printed itself: it did not. The staking register refreshed for the first time in two editions, and it came in stronger than the carried board. On the Ebunker series for the week ending October 11, total staked ETH stands at 43,865,190 (43.87M) — up from the 43.68M carried for two weeks — at a 36.34% staking ratio, with the network base APR at 2.51%, a +33bp step that restores most of September's 39bp fall. The stETH 7-day average APR printed 2.53%, and Lido's own October 6 statement put the stETH annualized yield at 2.23%, at or above its level since August. Read together, the network's willingness to keep capital at work barely moved while the token lost 6.9% in a week.

The queue remains the more interesting half. The exit queue rose to 964,365 ETH by the October 10 read (from 812,887 ETH on October 4) while the entry queue fell to 1,379,437 ETH (from 1,480,361), shrinking the buffer between them by 37.8% to 415,072 ETH. Demand to enter still exceeds desire to leave — a 3.2%-of-stake entry line against a 2.2% exit line — but the margin narrowed sharply. The churn arithmetic confirms both printed waits (16.7 days to exit, 23.9 days to enter, at the ~57,600 ETH/day ceiling). Active validators stood at 851,608, holding about 43.7M ETH, or roughly 35.8% of supply (Ebunker's 36.34% ratio uses its own denominator). The number that would change this reading is the buffer inverting; it did not.

Source: Ebunker Ethereum staking weekly (week ending Oct 11, 2026) for staked total, ratio, base APR and the stETH 7-day; Lido statement (Oct 6, 2026) for the 2.23% user-facing rate; validatorqueue.com via CryptoCompass and CryptoTicker (Oct 10, 2026) for queue levels, waits and validator count; churn arithmetic DERIVED.

2. stETH: supply and peg

The price had its widest week of the quarter and the peg still never broke. On the single-read basis stETH/ETH ran from a week-high +3.45bp (Oct 7) to a week-low −10.76bp (Oct 9, the liquidation session), closing at −1.85bp (Oct 11; ratio 0.99981) — a ±11bp band, narrower than last week's ±16bp, through a 9.8% ETH price range. A genuine redemption run prints discounts of tens to hundreds of basis points and stays there; this week's deepest print recovered within a day, and no queue formed on the stETH side at all. That is the market's verdict on whether the MetaMask episode or the price drawdown touched the protocol's solvency: neither did. Supply: still no fresh dated print — the 9,659,594 stETH figure is now over a month old, and the data report keeps the supply×price cross-check suspended. The substituted in-window derivation gives Lido 9,689,575 ETH under management (TVL ÷ price, Oct 11) and a 22.09% share of staked ETH (DERIVED against the Ebunker total) — down about 36bp from October 4, a move smaller than the TVL's because price enters both numerator and denominator.

Source: stETH/ETH series, DefiLlama coingecko feed, same-session prints; Lido TVL per DefiLlama protocol history (Oct 11 point); share-of-stake DERIVED on the TVL and Ebunker staked-total prints; supply per CryptoCompare via TickerReport (Sep 9, 2026).

3. DeFi / LSDfi

Fourth consecutive fully fresh board, and the first red majority in the record. Every liquid-staking row fell roughly 7–8% — Lido $24.30B (−8.3%), ether.fi Stake $4.78B (−7.7%), EigenCloud $6.52B (−7.4%), Binance Staked ETH $9.33B (−6.9%), Rocket Pool $1.31B (−6.8%) — tracking ETH's own price decline almost tick-for-tick, because their TVL is denominated in the asset that fell. The lending rows, holding a mixed book, fell less: Aave $18.46B (−3.3%), Morpho $11.28B (−0.3%), and Sky Lending $5.97B, the board's only riser, +1.4%. Ethereum chain TVL slipped −3.5% to $51.39B.

Two readings ride on the split. First, the LST rows are effectively a price chart this week, which is why Lido's −8.3% beating ETH's −6.9% is the one number with a non-price explanation: the MetaMask outflow layered a small supply reduction on top of the markdown. Second, and more usefully, the lending venues absorbed a $1.19 billion cross-asset liquidation cascade without a TVL collapse — a board that falls a third as much as the price it lends against is a board whose users are not fleeing, and Sky's small gain through the worst week of the quarter is a quiet confirmation that collateral, not confidence, is what these books hold.

Source: api.llama.fi protocol-history reads, one session, Oct 4 and Oct 11 points; Ethereum chain TVL per DefiLlama historical chain TVL.

4. Competition

The treasury bid named its own end. BitMine's October 7 hard cap on ETH supply at 5% puts it roughly 88,586 ETH from stopping; at its latest weekly pace (15,112 ETH in the week to October 4) that is about six weeks, though the company is not obliged to keep that rate. Its holding, 6,016,414 ETH, is about 4.9% of supply; roughly 84% of it is staked, which puts its staked balance near 11.6% of all staked ETH (DERIVED) — the largest single-operator slice this column tracks, and one whose yield behavior is now part of the network average's own path. The operator side of the competitive board printed two things worth naming. First, on October 5 Lido contributors proposed setting MetaMask Staking's target limit to zero in both the curated and community modules — no new deposits allocated to the operator — a governance action that closes the loop on the incident. Second, the published provider APRs: Lido's own stETH at 2.23% (Oct 6), the rETH mid-point carried at 2.19%, and the Lido user-net band unchanged at 2.21–2.24%.

Source: BitMine disclosure and Tom Lee remarks (Oct 7, 2026) via CoinPaprika and Coinspress; Lido governance proposal to zero MetaMask Staking's target limits (Oct 5, 2026) via the Lido research forum; Lido stETH APR per Lido (Oct 6, 2026); carried provider rows per their dated sources in the data report.

5. Outlook — and edition seven's four calls, graded

Edition seven published four watchlist items. One is settled cleanly, one is settled against us, one moves forward, and one stays pending.

Watchlist for the next edition, four items: (1) the buffer — entry minus exit at 415,072 ETH, down 37.8% in six days; whether it keeps shrinking or holds is the cleanest read on whether the repositioning has run its course; (2) the flow streak — five straight negative sessions for ETH ETFs; whether it breaks on a green day or a red one separates momentum from allocation; (3) the BitMine staked number — whether its staked balance moves off ~5.05M, since accumulation has flowed into the treasury rather than new validator positions; (4) the October 28 FOMC — the meeting that this week's minutes made, once again, a live question.

Source: scorecard inputs as cited in sections 1–4 and the data report.

6. The leverage layer, structurally

Worth separating what October 9 damaged, because the list is short and specific. Hit: roughly $1.19 billion of leveraged positions across crypto, about 89% of them longs; $356M of it ETH against $298M of bitcoin on a fifth of the market cap; nearly 193,000 traders; a single $19.98M ETH-USD position on Hyperliquid. ETH funding rates flipped negative for the first time in months. Untouched: the peg (inside ±11bp all week); stETH redemption; the protocol's staked balance (43.87M ETH, up on the refreshed board); staking yield (up 33bp); and Lido's share of stake (22.09% versus 22.45%, a move explained by the price that fell). The gap between those two lists is the whole week: a deleveraging event is a statement about who was borrowing against their conviction, not about the asset the conviction was in.

There is a durable lesson in the ratio. When a smaller-cap asset inside the same ecosystem takes a larger absolute liquidation total than the benchmark, the problem is structural positioning, not idiosyncratic news: traders had spent the $83,000–$87,000 bitcoin range building ETH exposure with insufficient margin, and the unwind fed itself once support broke. The same week, the protocol this site covers printed its yield up and its queue full but filling. Leverage re-forms after every cascade; the beacon chain's churn limit does not move.

Source: liquidation totals and the ETH/BTC ratio per BlockFence, Lumida News and CoinGlass (Oct 9, 2026); protocol data per the weekly data report.

7. The macro leg reversed

Last edition this column argued that the October 2 jobs report had, for the first time since the September hike, pointed the carry tables the right way without any staking yield changing. This week reversed that. The September FOMC minutes, released October 7, showed most officials expected another rate increase before the end of 2026; the 10-year Treasury touched a 24-year intraday high near 5.34% and settled in the 5.24–5.28% area, and the 30-year pushed toward 5.68%. On the refreshed 10-year print of 5.26%, the 10 ETH carry gap has widened to about −$762 a year, against −$426 versus the 3-month and −$532 versus the 2-year. For a yield-bearing asset, the macro leg is the competition, and this week the competition got harder, not easier. The October 28 meeting is once again a live question rather than a formality.

Source: FOMC September minutes (released Oct 7, 2026) and Treasury levels via BlockFence and the Gate Ventures weekly recap (Oct 5, 2026); carry gaps DERIVED on the Oct 11 ETH print.

8. Where Lido sits after the week

Score it structurally, because the headline number misleads. Lido TVL fell 8.3% to $24.30B — worse than ETH's 6.9% — but its share of staked ETH barely moved (22.45% → 22.09%, DERIVED), its network run by 600+ operators absorbed a large operator's full exit with the yield printing higher, and its peg held inside ±11bp through the worst price week of the quarter. The things that would genuinely dent Lido's position — a redemption failure, a peg break, a loss of user principal — did not happen, and this week tested that claim harder than September's event did, because the market itself was the aggressor. The carry-over risks are two and both observable: reward timing on MetaMask's returning stake (metered by the entry queue, now shorter), and the macro leg (metered by the October 28 meeting). Neither is a solvency question, and this week is the strongest evidence yet that those are the only two questions liquid staking leaves open.

Source: TVL, peg and share per the weekly data report; operator count per lido.fi (carried); incident figures per section 4 and the news review.

9. The trend this edition extends

Three threads this column has tracked for weeks all turned this edition, and the turns are what is worth recording. The staking register broke its two-edition silence and, in breaking it, resolved the only open question it had carried: 2.18% was noise. The queue's lead — entry over exit, the buffer this column introduced as the number that actually matters — shrank 37.8% in six days and is now the series to watch, because it moves faster than either queue alone. And the ETF flow series printed negative on all five sessions of the window, a run that now reaches back into the prior edition — across both a flat tape and a falling one — which is the definition of a flow series that is not chasing price. Against all three, the price did its own thing and the protocol ignored it. That divergence — a token down 6.7%, a yield up 33bp, a peg flat, a queue rebalancing — is the shape of a maturing asset, and it is the first time this site has been able to print it under genuine stress.

Source: series as cited in sections 1–8 and the data report.

FAQ

Related reading

• Lido & stETH Weekly Data Report — Week of October 5–11, 2026
• stETH Yield Update — Week of October 5–11, 2026
• The Steadiest Buyer: What BitMine's 5% Hard Cap Ends