The ETH-to-Stablecoin Carry Vault — A Case Study in Reading APR

A vault appeared advertising ~146% APR on ETH for what looked like a simple, elegant strategy. The structure is real and the protocols underneath are legitimate. The headline number is not. This is a case study in how to read a yield before you trust it — a skill that matters far more than this one vault.

The structure

The vault automates a four-step loop on Ethereum:

1. Deposit ETH as collateral.
2. Borrow a stablecoin against that ETH at a fixed, user-set interest rate, using a collateralized-debt protocol.
3. Provide that stablecoin as liquidity in a stablecoin pool to earn trading fees plus incentives.
4. Compound the rewards back into more ETH, and repeat.

On paper it's clean: borrow cheap, lend dear, pocket the spread ("carry"), and only ever touch two well-known protocols. That cleanliness is what makes it a good teaching example — because the math still doesn't work.

Why the headline number is wrong

Run the two legs at real, current rates:

Leg	Rate	Direction
Borrow the stablecoin (fixed CDP rate)	~5–7%	cost
LP the stablecoin in the pool	~3.1%	income

Unlevered, that's negative carry — you pay more to borrow than the pool pays you to lend. The only way to manufacture a triple-digit headline is to loop the leverage aggressively and catch a temporary incentive spike, then display the resulting peak as an annualized rate.

Two independent reality checks confirm the headline is inflated:

• The lending protocol's own public pool tracker lists this exact vault's real, revenue-grounded return at roughly 10% APR — about one-fourteenth of the advertised figure.
• The vault platform's own interface discloses that its dashboard performance figures are simulated, not realized.

Lesson 1 — Displayed APR is a marketing surface, not a return. Always reconstruct yield from its component legs. If the unlevered legs don't pencil, the headline is leverage, incentives, or a point-in-time snapshot — usually all three.

Where the real risk hides

The dangerous part isn't the modest real yield — it's what you take on to earn it.

• Forced unwinding ("redemption") risk. The borrow protocol lets any stablecoin holder redeem at face value against the borrower with the lowest interest rate first. A carry vault that sets the cheapest possible rate to widen its spread is, by design, first in line to have its ETH collateral pulled — without consent, even while the position is healthy. And the stablecoin trades slightly below $1 often enough that this is a live, ongoing event, not a tail case.

  Lesson 2 — A strategy optimized for the best entry price is often optimized for the worst exit. Read the mechanism that decides who gets hurt first, and check whether your strategy volunteers for that seat.

• Leverage cuts both ways. The same looping that lifts a ~10% real yield toward a triple-digit headline multiplies every downside move on the ETH collateral.

• Young, active-management contract risk. This vault type was only weeks old, and the platform had already suffered a ~$336K exploit earlier in the year in its active-management layer (not the simple vault shell). Newness plus a prior incident is a real cost, not a footnote.

• Thin exit liquidity. The stablecoin pool held only ~$7M. A large position dilutes the yield on the way in and pays slippage on the way out — and the vault uses queued (not instant) withdrawals, because positions must be unwound before you can leave.

• Incentive decay. Most of that ~3.1% pool yield is paid in incentive tokens, not organic trading fees. Organic fee-only yield is under 1%. Incentives shrink as more capital chases them and can be redirected at any time.

The verdict for a trader

Skip it as a yield play. The honest return (~10%) does not justify a leveraged ETH position that sits first in a forced-redemption queue, on a weeks-old vault, on a chain you may run nothing else on.

Keep the lesson. The underlying primitive — borrowing a stablecoin at a fixed rate against ETH — is a legitimate, well-audited tool. Used conservatively (a deliberately higher borrow rate to sit deep in the redemption queue, a comfortable collateral buffer, and no leverage loop), it's a reasonable way to get liquidity without selling. The vault's mistake is optimizing every dial for the headline instead of for survival.

The checklist this strategy teaches

Before trusting any advertised DeFi yield:

1. Rebuild the APR from its legs. If the unlevered spread is negative or thin, the headline is leverage/incentives/snapshot.
2. Separate organic yield from incentives. Ask what's left when the token rewards stop.
3. Find the "who gets hurt first" rule (liquidation order, redemption order, withdrawal queue) and check if your position volunteers for it.
4. Price the contract risk — age of the strategy, audit history, prior incidents, and how much power the manager holds.
5. Check exit liquidity and lockups before you need them, not after.

A high APR is a question, not an answer. This vault is the worked example.