Avant avETHx — an ETH label with a dollar filling
Data as of 31 Aug–1 Sep 2026 (Avant dashboard).
Bottom line: the junior tranche doesn't pay for its risk. avETHx runs 8.50x leverage — the highest of the junior tranches reviewed — with a reserve of $97,005 ahead of it: 0.313% of the product, covering just 2.66% of the tranche. The bigger point is elsewhere: 47% of this "ETH" product is the senior tranche of the same protocol's dollar product, which is three quarters USDe. ETH on the label, dollars inside.
What it is
The same three-layer structure as Avant's dollar product, but ETH-based: avETH is backed 1:1 by ETH, the senior savETH accrues yield as more ETH, and the junior avETHx absorbs losses after the reserve fund. Launched 18 September 2025 — under a year old at the time of writing.
Reserve as a share of product and junior leverage across three comparable structures. Dashed line — a cautious 1.5% threshold. Avant and Strata dashboards, 31 Aug 2026.
The numbers
| Item | Value | What it means |
|---|---|---|
| avETH product | $31.03m | NAV $31.34m, Pennyworks as of 25 Aug |
| Junior tranche | $3.65m | 11.76% of product → 8.50x leverage |
| Reserve fund | $97,005 | 0.313% of product · covers 2.66% of the tranche |
| Total protective buffer | $3.75m | 12.08% of product |
| Senior yield | 5.07% | |
| Junior yield | 7.75% | 1.53x |
The reserve can be checked without the dashboard: reading the reserve-fund address from the docs directly on-chain gave $94–103k; the dashboard shows $97,005. When the UI is silent, the address in the docs talks.
Better paid than the dollar product — still not much
The junior yields 7.75% versus 5.07% for the senior — 1.53x. That's noticeably more generous than Avant's dollar product, where the premium is just 1.23x. Realised history agrees: since launch on 18 Sep 2025 the senior has risen to 1.055 ETH and the junior to 1.088 ETH, +5.53% versus +8.81%, a ratio of 1.59.
The ETH tranche pays better for risk than Avant's dollar junior and worse than Strata's junior tranche on sUSDe.
ETH label, dollar filling
The asset list isn't topped by ETH:
Asset composition, Avant dashboard, 31 Aug 2026.
Avant's dollar product is 75.76% USDe. So through savUSD, the ETH junior tranche points at Ethena with roughly a third of its filling — on top of the ETH risk the name advertises.
That isn't diversification across products, it's entanglement inside one protocol. The protocol split confirms it: 15.04% of ETH-product assets are deployed into Avant itself. If the dollar product takes a loss, the ETH product inherits it automatically, with no separate event.
There's a reverse effect too — for savUSD holders. The ETH product holds $35.11m of savUSD. Against an estimated ~$102m for the whole senior dollar tranche, that's roughly a third of supply. So savUSD has a built-in largest holder: the protocol's own ETH product. In an ETH stress event, that holder sells savUSD. The senior dollar tranche size comes from a secondary source, so "a third" is an order of magnitude, not a precise figure.
Leverage revealed by the composition — lower than the dollar product
The composition percentages aren't computed off $31.03m: 47.10% equals $35.11m, so the base is $74.54m. Against net value of $31.34m that's a 2.38x loop. In the dollar product it's 7.6x — three times more modest here, which counts in the ETH product's favour.
It also explains the yield. ETH staking pays 2–2.76%. Multiply by 2.38 and you get 4.8–6.6%, and savETH's 5.07% lands right in that corridor. The senior tranche's premium over plain staking is leverage, not better strategies. That's an arithmetic plausibility check, not a protocol statement.
Half the assets sit on infrastructure without a track record
By chain: Ethereum 39.8% · Monad 35.8% · Sei 11.2% · Berachain 5.9% · Avalanche 4.8% · Movement 2.5%. 55.4% is on young chains, most of it on one. The dollar product has about 38% on young infrastructure.
By protocol: Aave 25.58%, Curvance 21.27%, Avant 15.04%, Morpho 14.35%, Spark 13.17%. Curvance takes a fifth here versus under 5% in the dollar product.
Senior savETH — a different answer from the dollar product
| Item | savETH | For comparison: savUSD |
|---|---|---|
| Tranche size | $27.33m | ~$102m |
| Buffer ahead of it | $3.75m = 13.7% | $18.83m = 18.5% |
| Yield | 5.07% in ETH | 10.47% in USD |
| Exit | 1 day | 1 day |
Versus the junior tranche the advantage is obvious: buffer rises from 0.313% to 13.7%, exit shrinks from a week to a day, and the cost is 2.68 points. But the ETH senior's buffer is thinner than the dollar senior's — 13.7% vs 18.5% — with the same leverage underneath. Tranche size reconciles: 10,494.93 savETH × $2,604.42 = $27.33m; plus the $3.65m junior that's $30.98m against a $31.03m product — there's practically no unstaked avETH.
For a directional holding, the premium drowns in ETH volatility
In the dollar product the senior pays 10.47% against a 3.83% risk-free rate: about six and a half points of premium on a base that barely moves. Here the arithmetic differs: the entire annual reward for protocol risk is erased by one ordinary daily move in the underlying.
For a hedged spot leg, the maths changes
If the ETH leg is hedged (spot against a short perp), volatility drops out of the equation: the premium stops drowning and becomes a clean add-on to the whole structure's yield. With funding around 10% annualised, an extra 2.31–3.07 points from the spot leg is roughly a quarter on top of the result — a lot.
But then two other things move to the front:
- exit liquidity — a hedged structure needs margin immediately on a sharp move, and one-day redemption is useless in stress, especially after the short leg is force-closed and the position suddenly becomes directional;
- collateral correlated with the income source — if 47% of the filling points, via the dollar tranche, at someone else's basis trade (Ethena), a drop in funding squeezes both the structure's income and its collateral's income at the same time.
Open question
47% of the filling is dollar savUSD, yet the tranche is ETH-denominated and has genuinely grown from 1.00 to 1.055 ETH. One of two things: either the dollar part is hedged back into ETH — adding hedge risk on top — or ETH exposure is incomplete and the holder owns less ETH than they think. The docs don't say.
Calls: one already refuted
A first estimate of this product (before dashboard data was available) made two calls. The dashboard checked both the same day:
| Call | Actual | |
|---|---|---|
| reserve stays below 0.5% of product | 0.313% | holding so far |
| savETH premium over staking won't exceed 2 points | +2.31…+3.07pp | refuted |
The error was in the inputs: a 4.5% savETH yield from a secondary source instead of 5.07% per the dashboard, and an inflated staking range (2.24–4.1% instead of 2–2.76%). It doesn't change the junior-tranche conclusion, which rests on 8.50x leverage over a 0.313% reserve, not on the premium.
What to watch
| When | What | Why it matters |
|---|---|---|
| monthly | reserve fund in absolute terms — dashboard or on-chain | not growing with the product → the ratio keeps falling |
| monthly | savUSD share of assets | below 25% — the entanglement weakens |
| quarterly | Monad share of chain allocation | above 40% on one young chain — infrastructure concentration |
Calls to check on 31 Oct 2026: the reserve stays below 0.5% of product; the savUSD share stays above 35%. If the second falls, the product really is turning towards ETH.
Unknowns
Obtainable:
- Twenty composition lines are hidden behind "show more" — six of twenty-six are visible. The share of actual ETH may be even lower.
- The exact size of the senior dollar tranche.
- Whether audits cover strategies, not just code.
- Two block explorers report different savETH supply — 10,494.93 vs 3,483; it doesn't affect the conclusions, since all calculations run off dollar values.
Unknowable from outside:
- How Monad behaves under forced-liquidation load — 35.8% of assets are there, and nobody has history.
- What happens to savUSD when its largest holder — the same protocol's ETH product — starts selling.
Who it suits, and when it breaks
On risk pay avETHx beats Avant's dollar junior (0.180 vs 0.165), the loop is three times more modest (2.38x vs 7.6x), and the protocol publishes its awkward numbers itself. The problem is the combination: the highest leverage of the three junior tranches sits on the thinnest buffer, and half the product isn't ETH. Anyone already exposed to Ethena adds to it again through avETHx.
For someone who wants to hold ETH with yield, plain staking pays 2–2.76% with no protocol layer, no junior tranche and no week-long cooldown. savETH is clearly better than the junior tranche, but for a directional holding it doesn't beat plain staking; for a hedged spot leg it's more interesting — provided exit liquidity and the collateral's correlation with funding are acceptable.
Sources: Avant dashboard — asset composition, chain and protocol split, Risk & Protection (31 Aug–1 Sep 2026) · Pennyworks NAV (25 Aug 2026) · Avant docs · on-chain read of the reserve fund · breakdowns of Ethena USDe and Strata jrUSDe (31 Aug 2026).
Research for information only. Not investment advice.