Lido & stETH Weekly Data Report — Week of September 14–20, 2026

1. Executive summary

The fifth snapshot is the first since edition one where the sector board is entirely fresh, and the first where a headline check fails. Six findings, none of them requiring an invented figure:

• Lido TVL rose 5.6% to $25.72B (DERIVED: 24.35 → 25.72, both direct API reads) — the largest one-week move since we began publishing, driven mostly by the price leg rather than new stake.
• The supply cross-check fails, at a 1.6% gap. 9,659,594 stETH × $2,621.87 = $25.33B against a TVL of $25.72B. We report the gap rather than move a number to close it, and section 3 explains why.
• Staking reached 35.77% on a circulating-supply denominator of 120.72M — the same denominator behind our Sep 7 print of 35.55%, making the two readings comparable for the first time in two weeks. The increase is +0.22pp.
• The exit queue stayed next to empty — 182,352 ETH with a 3.2-day wait on the Sep 20 read, against a 2,480,000 ETH entry queue running 43.1 days. That is a 13.6× imbalance, and it is a correction to what we published last week (section 6).
• LDO gained 12.7% in-window, its best week since August, on a supply basis that is now resolved: 834,151,322 tokens.
• The ETF bid turned negative for the first time in five weeks: −$140.6M net across the window's five sessions, after +$196.9M the week before.

2. Week-over-week table

Source: all figures as cited in sections 3–7 and the sources list. “Prior snapshot” values are the numbers this site published on Sep 15, kept for lineage even where the basis has since changed.

3. The cross-check, and why it fails this week

Our standing check multiplies stETH supply by the same-source stETH price and compares the product with Lido's TVL. It has passed twice. This week it does not:

The arithmetic is unambiguous: at today's price, a TVL of $25.72B implies about 9.81 million stETH in existence, roughly 150,000 more than the last supply figure we can date. The cause is staleness, not a broken model. Our supply print is eleven days old, and those eleven days covered in which the network's entry queue never fell below 1.75 million ETH. Aggregators disagree with each other, too: BingX prints 9,657,537 for Sep 7, which sits on our basis, while LBank prints 9.43M for Sep 18, which would widen the gap rather than close it.

Source: supply print as above; DefiLlama API (api.llama.fi/tvl/lido) and coingecko feed reads, Sep 20; aggregator prints via BingX and LBank price pages (accessed Sep 20).

4. The queues: 1.75M in, 182,352 out

The validator queues are the cleanest plumbing reading available this week, and they point one way:

The imbalance runs 13.6× in favour of entries (DERIVED). Ethereum rate-limits both directions to roughly 256 ETH per epoch, about 57,600 ETH a day; at that throughput the entry backlog is more than a month of work and the exit line clears in about 3.2 days. That asymmetry is the whole story of the 2026 queue: capital is willing to wait a month for 2.57%, and almost nobody wants their ETH back.

Source: validatorqueue.com reads as of Sep 13 (via CoinOKC) and Sep 20 (via Tremplin); churn limit per Ethereum protocol parameters as summarised by Crypto Briefing (Sep 2026).

5. The yield corridor against a new policy rate

With the Fed's decision in hand, the corridor can be printed against a policy rate that is no longer a forecast:

The gap arithmetic, on one price and one set of rates. Position: 10 ETH = $26,205.90 at the Sep 20 print of $2,620.59. Every line below is DERIVED:

Cash out-pays the carried user-net corridor by $446 a year against bills and $556 against the two-year. In basis points the shortfall runs 168–171bp against the 3M bill (DERIVED: 3.92 − 2.24 and 3.92 − 2.21) and 210–213bp against the two-year. Measured instead against the rate the Fed now pays banks on reserves, the staking shortfall is 133bp (DERIVED: 3.90 − 2.57 on the fresher network APR). The corridor's spread — frontend APR to net — is unchanged at 19bp.

Source: rates as tabled above; position value from the DefiLlama coingecko feed, Sep 20 print; all conversions are our own arithmetic on those inputs.

6. Data-quality ledger

Four items this week, three of them ours.

1. The exit-queue figures we published on Sep 15 were not from 2026. Edition four reported a queue “flip”: entry falling from 1.895M ETH (Sep 10) to 725,000 (Sep 12) while the exit queue went from 25,300 ETH to 2.656M ETH with a 46-day wait. That is not what happened in September 2026. The SwissBorg breach and Kiln's announcement that it would exit all of its Ethereum validators — the event our cited article attributes the queue to — occurred on Sep 8 and Sep 10, 2025, as SwissBorg's own incident report and its joint statement with Kiln both state. The September 2025 exit queue did reach roughly 2.6–2.67M ETH. Against that: validatorqueue reads 25,300 ETH on Sep 10, 2026, exactly zero on Sep 13, 2026, and 182,352 ETH on Sep 20, 2026. The protocol's churn limit processes about 57,600 ETH a day, so a 2.656M queue could not have drained between the Sep 12 and Sep 13 reads — and it did not, because it was never there. Our own same-day counter-read that week (exit zero, entry 1,858,000) was closer to the truth and we set it aside. Corrections are published above the affected editions.

2. The peg series changes basis. The 0.9987 we printed on Sep 15 was a live same-timestamp pair read at the snapshot moment — honest, but not reproducible. From this snapshot the ratio is computed from the daily series' same-source same-day prints, so any reader can re-derive it. On that basis Sep 15 reads 0.99969 (−3bp) rather than 0.9987 (−13bp): a 9bp difference in basis, not a revision of the market. A basis change is disclosed here because it changes the series, not because the old print was wrong.

3. The LDO supply basis is resolved — downward. Four editions carried 895.8M LDO outstanding from a tracker we could not corroborate. A Sep 17 market data set puts circulating supply at 834,151,322 against a 1B total. That is 6.9% below our carried basis, which means every LDO market cap we have printed since August was roughly 7% too high. This edition uses 834.151M; the old basis would produce $375M today against the $349M we now print.

4. The sector board is fresh for the first time since edition one. Every row is a direct API read taken Sep 20 — no reprints. One consequence: the “share of cross-chain” figure on our homepage card is retired, because the chain endpoint does not return a total we can defend. The card now carries a week-over-week change instead.

Source: SwissBorg incident report and Kiln joint statement (both Sep 2025, retrieved Sep 20); validatorqueue reads as cited in section 4; circulating-supply figure per Analytics Insight (Sep 18, 2026); DefiLlama API reads, Sep 20.

7. Macro and flows

The policy turn. The FOMC raised the target range to 3.75–4.00% on Sep 16 in a 12–0 vote, the first increase since July 2023, and its median projection holds the funds rate at 4.1% through this year and next. Total PCE is projected at 3.7% for 2026, easing to 2.3% in 2027; the unemployment rate is put at 4.1%.

ETF flows flipped negative. From the Farside daily table: +$121.1M (Sep 14), −$142.0M (Sep 15), −$224.1M (Sep 16), −$39.3M (Sep 17), +$143.7M (Sep 18). Window net: −$140.6M, DERIVED. The two worst sessions were the Fed statement day and the day after; the two best bracketed them.

The sector board, refreshed. Every row below is a DefiLlama direct read taken Sep 20:

Two ratios fall out of that board. Lido is 2.62× the size of the largest custodial liquid staking token (Binance Staked ETH) and 1.35× the largest DeFi protocol on any chain (Aave) — DERIVED, both. The four liquid-staking protocols on this board hold $41.81B between them, and Lido plus Binance Staked ETH account for $35.53B of it — 85% (DERIVED).

Source: FOMC statement and SEP, Sep 16; Farside Investors daily table, accessed Sep 20; DefiLlama API direct reads (api.llama.fi/tvl/<protocol> and /v2/chains), Sep 20.

8. Conclusion

The week's data says two things at once, and they are not in conflict. The tradable layer repriced violently — ETH 4.9% down on the Fed, then 9.2% up off the low; ETF flows negative on the week; LDO swinging 10% down and 25% up inside seven sessions. The structural layer did not move at all: the entry queue held 1.75M ETH, the exit queue held 182,352, the staking ratio rose 0.22pp and the network's validator count actually fell slightly as consolidation began. The fifth snapshot's headline failure is a stale supply print, not a broken market, and the honest thing to do with a 1.6% gap is to leave it standing and say why.

FAQ

Related reading

• Lido & stETH Weekly News Review — Week of September 14–20, 2026
• 8.4M ETH on the Move: What Lido's CMv2 Consolidation Actually Changes
• Lido & stETH Weekly Data Report — Week of September 7–13, 2026