Ethena Pay's Earn+ Puts Four Companies and a 24-Hour Exit Behind One Button

One tap. That's all a user sees. Underneath that tap sits an Upshift vault, a curator called Steakhouse, a strategy that supplies one stablecoin and borrows a different one, and Aave V4 running on Avalanche. Four companies, one screen, and a 24-hour window to get out.

That's the actual product. Not a savings account. A carry trade wearing a consumer interface.

Four Names, One Button

On October 28, Aave Labs published the details of how Earn+ works inside Ethena Pay. Deposits route into an Upshift vault. Steakhouse curates that vault. The strategy supplies USDe and borrows USDT. The lending market underneath is Aave V4 on Avalanche, which shipped there in July with the Hub and Spoke design.

Here's the sentence that matters most. Aave says plainly that it doesn't operate, curate, or manage the vault.

Read that twice. The name with the reputation is providing rails. The judgment calls happen one layer up, with a curator that almost no retail user could have named last week.

Withdrawals clear inside 24 hours per the stated terms. Not instantly. A window. And users pick how much of their balance to allocate, which means the risk sizing is self-service.

Is any of this dishonest? No. It's disclosed. But disclosure and comprehension are different things, and a familiar app design tends to flatten them.

The Borrow Side Nobody Prices

Everyone talks about the yield. Almost nobody talks about the funding cost that sits on the other leg, and that's the whole ballgame.

A position that supplies USDe and borrows USDT earns on one side and pays on the other. Net rewards equal the supply yield minus the borrow rate, minus fees, minus slippage whenever the vault has to rebalance. If USDT borrowing gets expensive, the advertised number compresses and the collateral doesn't have to move a cent for that to happen.

But the interface shows one number. A rate. Printed like it's a fact about the product rather than a snapshot of a spread that moves every block.

The consensus trade is crowded. Everyone in this sector has spent two years assuming funding stays positive and borrow stays cheap because that's what the last two years looked like. I've seen this movie before. The version where the spread inverts and everyone discovers they were running a levered position, not a savings account.

So here's my first take. The interesting story on October 28 wasn't the yield. It was Aave becoming invisible infrastructure. A consumer fintech app just embedded a lending protocol as a backend component, the same way consumer apps embed payments processors. Aave gets volume and distribution without the curation liability. Ethena gets another demand sink for USDe. Upshift and Steakhouse get assets. All four win on the way in.

And on the way out? Different question entirely.

Which brings me to the second take, and it's the one that'll annoy people. The 24-hour processing window is the most honest thing in the entire announcement, and it's also the part users will ignore. A window exists because unwinding a borrow position and retrieving assets takes time. That's not a flaw. It's a tell. It tells you the underlying exposure isn't liquid in the way a checking balance is liquid.

What if the opposite is true? What if the smooth app experience is actually a risk multiplier? The easier something is to enter, the more people enter without reading, and the more crowded the exit gets when the rate drops. Sentiment extremes don't announce themselves. They look like a nice APR and a green button.

When the crowd panics, I sharpen my pencil. But I'd rather sharpen it before the panic, which means asking who's actually on the hook if a layer breaks. Collateral valuation. Liquidation parameters. Smart contract behavior. Curator discretion. Four separate failure points, four separate parties, and no single entity accountable for the combined outcome.

What I'd Actually Watch

Ignore the headline rate. Watch the USDT borrow cost on Aave's Avalanche market. That's the number that decides whether this product keeps paying or quietly stops.

Then watch the curator. Steakhouse can change the strategy. Users can't. That asymmetry is the real risk surface, and it doesn't show up anywhere on a dashboard.

Earn+ isn't a scam and it isn't a free lunch. It's a packaging decision, and packaging is fine as long as you know what's inside the box. The people who get hurt in structures like this aren't the ones who took risk. They're the ones who didn't know they had.

So treat it as a net-yield question, not a savings label. Subtract the borrow cost. Subtract the fees. Ask how fast the vault can actually unwind. If the answer still looks good, fine, allocate. If the answer depends on funding staying where it's, you're not earning yield. You're short volatility and getting paid for it until you're not.