Ethena held through the run on redemptions and came out a third of the size

How does a synthetic dollar stay still?

The arithmetic is elementary. If the collateral rises ten percent, the bought position gains ten and the sold one loses the same. The sum stays zero, and that is what lets Ethena back every dollar issued with one dollar only, where an overcollateralized stablecoin like DAI asks for one and a half.

The balance rests on one assumption, though: that the spot price and the perpetual’s move together. In normal conditions it is a solid one, because arbitrageurs keep them aligned. When an exchange suspends trading or the liquidity evaporates, the two prices come apart — and in those hours the hedge doesn’t hedge.

How much real adoption is left?

It isn’t a run. It is leverage deflating. What left were the positions built on top of the yield — deposit, borrow, deposit again, start over — not the wallets holding USDe as a dollar. sUSDe’s yield, meanwhile, is at 11%.

What is left is still scaffolding that a good part of decentralized finance has come to lean on. USDe is accepted as collateral by Aave, Pendle, Morpho and Curve, and unpicking those integrations costs more than keeping them. A protocol deflating is a different thing from one disappearing.

Does the revenue hold when the market cools?

It is there to keep the perpetual hooked to the spot price. When many people buy with leverage the perpetual trades above spot, and every eight hours the longs pay the shorts. Ethena, always on the side that sells short, collects. When the optimism deflates the payment compresses, and the revenue with it.

Underneath that payment there is a base that doesn’t depend on the market’s mood, though: the share of collateral held in staked ether pays three or four points a year regardless. It is why in June sUSDe’s yield fell to 7.1 percent without ever going to zero.

How liquid is USDe on the day it counts?

On 10 October 2025 USDe touched $0.65 on Binance’s order book during the liquidation cascade, while remaining entirely backed; the price came back within hours.

Backed and liquid aren’t the same thing. The collateral was all there, but at that moment whoever could have bought was closing their own positions, and the price went where the few orders left sent it. Whoever sold there genuinely lost; whoever waited saw the dollar again.

This risk can’t be removed, because it doesn’t depend on Ethena but on the market where USDe trades. It can be accounted for, though: if you need money available at a moment’s notice, USDe isn’t the place.

How many counterparties does the collateral depend on?

Keeping them separate is the most intelligent thing they have done. Ethena doesn’t transfer the collateral to the exchanges. It pledges it as margin, and legal ownership stays with the custodian. If an exchange fails, that collateral doesn’t end up in the bankruptcy estate.

We already know it works. In February 2025 Bybit, one of the exchanges where Ethena held its hedge, suffered the largest theft in the sector’s history — about a billion and a half dollars. Ethena’s collateral was at the custodian, and only some thirty million of derivatives margin was left on that exchange, taken to zero the same day. Ethena lost nothing.

A real dependency remains, though. In a systemic crisis the trouble arrives on several exchanges at once, and in that case the hedge would become a debt to collect instead of a position to close.

Does the governance count for anything?

The reason the fee switch doesn’t get turned on isn’t technical, it is a real conflict. Diverting a share of the revenue to whoever holds the token means taking it away from sUSDe’s yield, which is to say weakening the only reason anyone prefers USDe to any other dollar. It is the same money and the two sides are fighting over it. So far the side that brings customers has won.

This is a judgment and not a measurement, because there is no sensible way to measure how real a governance is. What can be seen is that the important decisions — which exchanges to be on, how much to keep in the reserve fund, when to change collateral — have never gone through a vote.

Does the team hold up?

The tests come through: in February 2025 the billion-and-a-half theft suffered by Bybit, with Ethena exposed for some thirty million on derivatives and no losses; on 10 October 2025 the liquidation cascade, with USDe at $0.65 and the price back; between October 2025 and March 2026 the contraction from fourteen to $5.92 billion, declared in the quarterly reports.

And it told the story well. The fall from fourteen to five billion is in the quarterly reports, without circumlocutions; so is the drop in revenue; and they wrote about October’s depeg instead of pretending nothing had happened. In a sector where the uncomfortable numbers disappear from the press releases, that is not nothing.

It is a judgment, so it holds as long as the people who earned it stay. And in a system where the hedges have to be rebalanced across four exchanges twenty-four hours a day, the team and the product are the same thing.

How much does the mechanism risk?

Against the first scenario there is a declared defense: the protocol keeps five percent of that payment when it goes its way and sets it aside in a reserve fund. On the history available it holds — even in the worst years the days when the payment went the other way were about one in ten. But that fund is calibrated on episodes of days, not months.

Put bluntly: every time a stablecoin offers a double-digit yield, the comparison with Luna and UST arrives before any analysis of the mechanism.

The difference from a year ago is that two scenarios have left the theory. The cascade arrived and the price came back; the attack on an exchange arrived and the exposure stayed marginal. What nobody has seen yet is the long scenario, months of negative payments in a row — the one case the reserve fund can’t cover.

Comparison between ethena and the direct competitors in the same category, on the measured criteria and all taken on 2026-08-20 from the same source (defillama). In circulation: ethena $4.08bn, sky $6.69bn, usual $550m. Capital backing: ethena $4.08bn, sky $5.62bn, usual $90m. Revenue over 30 days: ethena $0.03m, sky $13.8m, usual $0.39m.