Aave's USDe Rate Hike Could Flip Ethena's Carry Trade Negative

What happens when the cost of borrowing USDe climbs higher than what sUSDe actually pays out?

That's the question sitting on every Ethena loop trader's desk right now. And the answer is uncomfortable. LlamaRisk wants to raise borrowing costs on USDe across five Aave V3 markets, and the math on the popular carry trade gets ugly fast.

The Raw Numbers

Let me break this down. The Sept. 9 recommendation pushes USDe's base variable borrow rate from 5% to 6% on Core, Plasma, Monad, Mantle, and Avalanche. Slope1 comes down by one percentage point at the same time. That combination adds 13 to 89 basis points of cost across a $323.8 million debt snapshot.

Not every market gets hit the same. The spread depends on where utilization sits and how much of the curve gets repriced. But the direction only goes one way. Borrowing USDe gets more expensive.

The numbers tell the story. A loop that earned a comfortable spread a few weeks ago is staring at thin margins now, and in some configurations, a negative one.

Why the Loop Breaks

Here's the trade everyone ran. Borrow USDe on Aave, swap it, deposit into sUSDe, collect the yield, pocket the difference. It works as long as sUSDe's yield clears the borrow rate plus whatever slippage and gas you burn on the way in.

Ethena's sUSDe yield has been compressing as funding rates on perps cool off. That's the funding source. When funding is rich, sUSDe pays a lot and the loop prints. When funding flattens, the whole structure gets squeezed from both sides.

So you've got rising borrow costs on one end and falling collateral yield on the other. That's the definition of negative carry. And notably, it doesn't take much. A 13 basis point bump kills the marginal position. An 89 basis point bump kills a lot more.

From a risk perspective, this is LlamaRisk doing exactly what it's supposed to do. USDe is a synthetic dollar with its own set of dependencies, and letting borrowed positions stack on top of it at subsidized rates is how you get a cascade when something finally breaks. Higher borrow costs cool the loop before the market does it for them.

What the Desk Thinks

Traders are watching the governance thread closely. According to the proposal, the changes target utilization and interest rate model parameters together, which tells you the goal isn't just revenue. It's risk containment.

The street's read is straightforward. Loops unwind when carry goes negative, and unwinds are mechanical. You repay USDe, you redeem sUSDe, you sell the underlying. Do that across $323.8 million and you get real selling pressure in spots nobody's watching.

What the street is missing: the unwind doesn't have to be violent to matter. Slow bleed is worse for positioning because nobody panics, they just quietly close positions, and the yield that attracted capital in the first place keeps drifting lower.

What to Watch

The governance vote is the first catalyst. If it passes, the new curve goes live and the marginal loop's economics reset within days.

Watch sUSDe's yield next. If it holds above 7%, most loops survive the hike. If it slips into the 5s, the negative carry flips from theoretical to real.

And keep an eye on Aave's USDe utilization rate. A drop there tells you the unwind has started. A spike tells you someone's still chasing the trade and hasn't done the math.

The carry trade isn't dead. But it's a lot less forgiving than it was two weeks ago.