ether.fi weETH after restaking — zero premium over Lido and a liquidity wall
Data as of 8 Sep 2026.
Bottom line: weETH is no better than wstETH, and worse where it matters for collateral. On 7 August 2026 ether.fi pulled restaking out of weETH — it's now ordinary staked ETH competing with Lido. Its premium over wstETH, measured directly on-chain over a year, is zero. The protocol's redemption works and has passed a real stress test. But market exit barely exists: only 0.63% of supply can be sold without heavy slippage, and 67.8% of all weETH sits as collateral in a single Aave contract.
What it is
weETH is a liquid staking wrapper from ether.fi: you deposit ETH and receive a token whose exchange rate slowly accretes against the underlying from validator rewards.
Until 7 August 2026 it was a "restaking" token: the same ETH also served as collateral for third-party services via EigenLayer and was supposed to earn a premium for it, in exchange for extra slashing risk. That's gone. ether.fi moved restaking into a separate token, weETHs, and left weETH as plain staked ETH. ether.fi's CEO called it the end of an era and said he sees restaking as premature rather than dead. Per the protocol's docs, The Defiant and CoinDesk (7 Aug 2026), less than 1% of assets remain in restaking, versus about half at the start of 2026.
No premium over Lido
Measured the cleanest way available: the share rate read from the contracts at historical blocks. Market price drags discounts and premia along; the share rate doesn't. Three independent archive nodes returned identical values.
| Period | weETH | wstETH (Lido) | rETH (Rocket) | weETH − wstETH |
|---|---|---|---|---|
| 30 days | 2.39% | 2.24% | 2.19% | +15bp |
| 90 days | 2.42% | 2.27% | 2.19% | +15bp |
| 180 days | 2.46% | 2.40% | 2.13% | +6bp |
| 365 days | 2.50% | 2.50% | 2.26% | 0bp |
The measurement floor. wstETH and rETH are two plain staked-ETH tokens that should earn almost the same by design. The gap between them over the same windows is 5, 8, 27 and 23bp. weETH's edge (0–15bp) is smaller than the gap between two tokens that shouldn't differ at all. That's not a spread, it's noise. ether.fi and Lido charge the same 10% protocol fee. weETH currently has no points, airdrops or emissions: season 5 ended on 31 May 2025.
Redemption exists, works, and survived stress
A staking wrapper without direct redemption is structurally the same as the one that broke in April 2024 (ezETH, a $60–65m liquidation cascade). weETH has redemption: unwrap → queue → receive ETH.
- Stated time — about 10 days.
- Observed under load: median 4.9 days, max 16.7 days.
- Ethereum's validator exit queue at the time — 96 ETH, about 2 minutes. The entry queue — 1.95m ETH and almost 34 days: people want in, not out.
The stress test is independently confirmed. ether.fi reported redeeming 542,792 ETH — 19.6% of all assets — between 18 April and 21 May 2026, over 33 days, without a single failed payout. Reading the chain over the same period shows a drop of 654,660 ETH in 31 days — same event, same order of magnitude. The protocol's claim about itself held up to outside measurement.
Money left before restaking was removed — and is coming back after
Eleven points taken from the weETH contract at historical blocks on 8 Sep 2026: token supply times the share rate at the same block. From the January 2026 peak to the June low — −46.6%. From the low to the time of writing — +24.8%, and the growth continued after restaking was removed.
The outflow happened before restaking was pulled and has already reversed. ether.fi's decision looks like a response to capital flight, not its cause.
But "coming back" doesn't mean what it seems. Roughly four fifths of new inflows went straight into Aave (145k of 174k tokens over 25 days). What's returning isn't holders, it's demand to borrow against weETH: growth in size and growth in leverage on top of the asset are the same growth. This is a 25-day balance difference from two sources, not a traced flow, so the defensible range is 80–85%.
You can sell 0.63% of supply. Then there's a wall
Measured with router quotes on 8 Sep 2026. Up to 12,000 weETH, slippage is 0.7%. Between 12,200 and 15,000 it jumps from 1.5% to 20.5%. After that proceeds stop rising altogether: 100,000 tokens fetch the same as 30,000. Liquidity sits tightly around the rate and vanishes as soon as price is pushed out of range.
In total about $33m can be sold on venues against $5.3bn of supply — 0.63%. CEXs don't help: three venues, $6.9k of daily volume.
The same measurement for wstETH, same tool, same minute:
| weETH | wstETH (Lido) | |
|---|---|---|
| Sell depth within 2% slippage | ~$33.4m | ~$104.2m |
| Same, as share of supply | 0.63% | 0.91% |
| Daily volume | $1.63m | $4.55m |
| Cliff starts after | 12,000 tokens | 33,000 tokens |
wstETH is 3.1x deeper in absolute terms but only 1.4x deeper per dollar of supply. And both have a cliff: wstETH drops from −0.9% to −12.9% between 33,000 and 40,000 tokens. So wstETH is 1.4x better here, not safer in kind — and neither could absorb an Aave-scale unwind.
What actually separates weETH from Lido is the direction of capital: over the year weETH supply shrank 23.6% while wstETH supply grew 18.5%.
The structural takeaway. There's practically no market exit from weETH — the real exit is the protocol's redemption queue. The April–May event proves it: 650k ETH went through the queue because the pools couldn't have absorbed even one percent. For anyone who can wait ten days, that's fine. For anyone levered, it isn't: a liquidation doesn't wait ten days, it sells into that same $33m pool.
Two thirds of supply in one contract
| Holder | Share |
|---|---|
| Aave (lending collateral) | 67.78% |
| LayerZero bridge | 5.70% |
| Morpho (lending) | 5.43% |
| Arbitrum gateway | 2.59% |
| Spark (lending) | 1.98% |
| Unidentified | 1.98% |
| Top 4 combined | 81.50% |
| In lending combined | 75.19% |
Aave raised its weETH supply cap on 14 Aug 2026 from 1,220,000 to 1,350,000 — and within three and a half weeks it was 97.1% full. Demand for weETH today is mostly demand to borrow against it, not to hold it.
Galaxy Research quantifies the consequence: fewer than 9% of Aave loans carry about half of all debt at an average health factor of 1.06, and weETH is about 42% of that group's collateral. An 8–9% discount triggers mass liquidations; a modelled 10% gap leaves $2.47bn of debt against $2.42bn of collateral after the shock. Two outlets cite these numbers, but both point to the same study — one source repeated twice.
Pendle no longer exists for weETH
In March 2026 Pendle — the main venue where holders of such wrappers took leverage — lost 87.6% of revenue ($4.44m a month in August 2025 versus $552k in March 2026). Active weETH markets on Pendle at the time of writing: zero. If anything that removes a risk (no term wrapper forcing all exits onto one date), but it's another sign the infrastructure around the asset has dried up.
As basis-trade collateral: weETH vs wstETH
The typical market use for tokens like this is the staked part of the spot leg in a basis trade. In a hedged leg ETH volatility drops out, and three things decide — yield, exit liquidity and collateral correlation. On all three weETH either matches wstETH or loses:
- yield: 0–15bp of edge against a 5–27bp measurement floor — i.e. zero;
- exit: $33m vs $104m — 1.4x worse per dollar of supply, with a cliff in both; the real exit is a 5–17-day queue;
- collateral correlation: 67.8% of supply in one Aave contract, with a 1.06 average health factor in the cohort carrying half the debt. That's other people's leverage, but it sets the price at which everyone exits.
Net of the comparison: in this role weETH earns no premium over wstETH and is weaker on exit and collateral concentration. It suits those who can wait out a 5–17-day redemption queue. It breaks when exit has to go through the market rather than the queue: a liquidation sells into the same ~$33m of depth.
What would change the view
Call 1. By 8 Dec 2026, the realised weETH-minus-wstETH spread on 90-day share-rate growth won't exceed 30bp. If it does, the mechanism is different from the one described.
Call 2. By 8 Dec 2026, weETH's ETH-equivalent size won't fall below 1.71m ETH — the June low. A break means the outflow has a different cause that's still active.
Downside. If Aave raises the cap again and it fills within a month while pool depth doesn't grow, leverage on the asset rises and exit doesn't. Then the 8–9% discount in Galaxy's model becomes a matter of time.
Upside. Worth revisiting only if, at the same time, the realised spread over wstETH stays above 50bp for two straight quarters and sell depth within 2% slippage exceeds $100m. One without the other doesn't count — premium without exit has already ended in liquidations once in this asset class.
Unknowns
Obtainable:
- Historical pool depth (paid access). The proxy is volume: weETH's fell from $11.2m to $1.63m a day over the year, 7x, while wstETH's fell 8.1x. The whole category is drying up, Lido even faster — so liquidity is a category-wide problem, not a weETH-only one. That's volume, not depth.
- Worst historical discount: about −1.2% in Jan–Feb 2024 and about −0.5% in April 2026 — two points of differing reliability. The discount is what triggers the cascade in Galaxy's model, so this is the costliest gap.
- Depth was measured with one router at one moment.
Unknowable from outside:
- Whether the redemption queue can handle an exit larger than April's. The largest observed stress was 19.6% of assets over 33 days; nobody knows what 40% looks like.
- Whether restaking returns, and on what terms.
Sources: direct reads of weETH, wstETH and rETH contracts at archive blocks (8 Sep 2026, three independent nodes) · ether.fi docs · The Defiant, CoinDesk (7 Aug 2026) · router quotes (8 Sep 2026) · Etherscan, weETH holders · Aave supply caps · Galaxy Research · DefiLlama · StakingRewards.
Research for information only. Not investment advice.