Lido & stETH Weekly Data Report — Week of September 7–13, 2026

1. Executive summary

The fourth snapshot is the mirror image of the third: the market barely moved, and the plumbing flipped. Five findings, none of which required us to invent anything:

• The exit queue is no longer empty. From exactly zero ETH (Sep 2 read) to 2.656M ETH with a 46-day wait (Sep 12), driven by Kiln's orderly exit of ~1.6M ETH after the SwissBorg hack.
• The entry queue collapsed by 65% — from 2.074M ETH / 36 days to 725,000 ETH / 12.6 days (DERIVED: 0.725 / 2.074270 − 1 = −65%).
• The cross-check passes again. 9,659,594 stETH × $2,514.60 = $24.29B against a DefiLlama TVL of $24.35B — a 0.2% gap, our second consecutive pass on a corrected basis.
• stETH supply dipped, slightly. 9,661,935 → 9,659,594 (−2,341 stETH, −0.02%) — flat-to-down supply in a record-staking week, with a 10,000 ETH redemption in the flow.
• LDO gave back 8–9% of last week's +11% — with no Lido-specific disclosure in the window to attribute it to.

One housekeeping item runs through this report: our ETH and LDO daily series changed basis (see section 6). Every value below carries its date and its source, as always.

2. The week-over-week table — 16 metrics, each with its date

3. The cross-check — supply × price vs TVL, on one date

Our founding methodology requires stETH supply × price to reconcile with Lido TVL. On the Sep 15 reads: 9,659,594 stETH × $2,514.60 = $24.29B against a DefiLlama TVL of $24.35B. The gap is 0.2% (DERIVED: 24.35 / 24.29 − 1) — inside rounding, and the second consecutive pass since we corrected the mispriced August TVL. Two caveats, stated rather than buried: the price and TVL are same-source-family reads (DefiLlama's coingecko feed and DefiLlama's protocol TVL, both Sep 15), and the supply print is Sep 9 — six days earlier, from CryptoCompare. A 0.2% gap on inputs that are not all same-day is a pass with a timestamp, not a pass forever.

4. The queue flip, quantified

Three dated reads make the week's shape:

• Sep 2 (prior edition): entry 2,074,270 ETH / 36 days; exit zero.
• Sep 10: entry 1.895M ETH / ~33 days; exit 25,300 ETH / ~10.5 hours.
• Sep 12: entry 725,000 ETH / 12d14h; exit 2,656,000 ETH / 46d3h.

At our Sep 12 series price of $2,513.56, the exiting pile is ~$6.68B (DERIVED) — against the BlockBeats article's own $9.03B figure, which we reject because it implies $3,400/ETH. The driver is dated and named: Kiln's Sep 10 announcement of an orderly exit across its ~1.6M ETH of delegated stake. If Kiln's re-stake intention holds, the natural next state is the exit queue draining back toward zero while the entry queue re-fills by roughly the same amount — the scenario math is in this week's deep dive.

Two asymmetries in the queue reads deserve printing beside the raw numbers. First, the wait asymmetry: 46 days to exit against 12.6 to enter is a 3.7-to-1 ratio (DERIVED) — the machinery that admitted deposits freely for weeks now charges 3.7 times longer to leave, which is how the protocol prices patience without forbidding exit. Second, the drain-rate yardstick: at steady state the exit queue implies roughly 57,700 ETH leaving per day (DERIVED: 2.656M / 46) — about 1,800 validators a day at 32 ETH, the same daily figure last week's report computed for entries. Same pipe, opposite direction, same arithmetic; next week's print will show whether reality matches it.

5. Concentration: 5.067M staked ETH on one balance sheet

BitMine's Sep 8 disclosure puts 5.067M ETH of its 5.93M treasury in staked positions. Against the Sep 12 network total of 43.1M staked, that is 11.8% (DERIVED) — and the full treasury is ~4.9% of circulating supply. The same disclosure added 28,086 ETH in the week, about 0.55% of its own staked stack (DERIVED), and roughly 16% of the network's ~180,000 ETH staked growth on the week (DERIVED: 28,086 / 180,000). We flag concentration as data, not as judgment: the network's validator count keeps rising (910.9k), which is the other half of the concentration story — more validators, but a growing share of their delegated stake traceable to fewer treasuries.

6. Data-quality ledger — including one series change

• Price-series change (ours). CoinGecko's daily-close page is no longer reachable to our pipeline, so the ETH and LDO daily series move to DefiLlama's coingecko feed (one print per day at a fixed UTC hour). Basis check: the Sep 6 equivalent reads $2,507.25 on the new series vs $2,479.92 as the prior close print — a 1.1% gap between bases; the LDO Sep 8 equivalent ($0.39583) sits 1.1% above CoinLore's same-day $0.3917. Prior prints remain in lineage on the archived pages.
• Staking-ratio denominators differ. 35.55% (Sep 7) implies ~120.7M circulating; 35.29% (Sep 12) implies ~122.1M. Both prints are quoted verbatim; the ratio row is marked not-comparable rather than averaged.
• Two contradictory Sep 12 queue reads. BlockBeats (exit 2.656M / entry 725k) vs a staking-survey counter-read (exit zero / entry 1.858M). We carry the read consistent with the named Sep 10 Kiln event and print the conflict. We also rejected the BlockBeats dollar conversions (implied $3,400/ETH) while keeping its ETH counts and wait times.
• LDO supply basis: fourth week unresolved. CoinLore 895.8M vs CandleRank/Hexn 836.3M vs a third implied basis. Market cap is printed on the CoinLore basis and labelled as such.
• No fresh provider APR prints. The stETH/rETH rate board is carried (Sep 2/Sep 7 reads) and labelled; the fresh network-level APR is 2.59% (Sep 10).

7. Macro and flows

The 3-month bill stays at the carried 3.92% print (Sep 1) — against a 2.59% network staking APR (Sep 10), the gap is 133bp (DERIVED), one basis point wider than last week's 131bp. The FOMC meets September 15–16, one day after this snapshot; our last carried odds print was 56% (Aug 28). The ETF tape is in section 2's row 14 and the news review's item 7. Bitcoin's carried print ($79,992, Sep 6) had no fresh in-window read we could verify, so it is not re-quoted.

The FOMC translation table, printed before the meeting so next week can grade against it: the carried 3.92% bill sets a 133bp shortfall against the 2.59% network APR (DERIVED). Each 25bp the committee delivers moves the gap 25bp toward the staker (DERIVED); with the last odds print at 56% (Aug 28), much of the hike path already sits inside current yields, so the grading question is the deviation from those odds, not the direction. On the flow side, the tape's four sessions (-24.3, +34.7, -29.9, +216.4 $M) net to +$196.9M — and the Thursday print exceeds the other three combined (net -$19.5M) by an order of magnitude (DERIVED: 216.4 / 19.5). Concentration of flows inside the window matters as much as their sum.

FAQ

Related reading

• The Queue Flips: 2.66M ETH Is Now Waiting 46 Days to Leave
• Lido & stETH Weekly News Review — Week of September 7–13, 2026