What is the cheapest USDC borrow rate right now?

This was typed into the product word for word. It is a good question with an annoying answer: there is no single cheapest rate, and the thing that decides what you actually pay is usually not the one you are comparing.

Start with the size of the spread#

Rates are not close. Two venues on the same asset, on the same chain, at the same moment routinely differ by more than a percentage point.

That is not a rounding difference. On $100,000 borrowed it is over $1,000 a year, and it exists between venues most people treat as interchangeable.

Three networks each, so “the cheapest USDC borrow rate” is really a question about a dozen or so numbers that move independently.

The advertised rate is a snapshot of a moving thing#

Four numbers matter, and most comparisons only show you one.

1. Borrow APY

   What you are quoted. The one everybody compares.

2. Utilisation — how much is already borrowed

   This is the one that decides where the rate goes next. Lending markets price by utilisation: as more of a pool is borrowed, the rate climbs, often sharply past a kink. A cheap rate at 40% utilisation and a cheap rate at 90% are different propositions — the second is one large borrower away from not being cheap.

3. The loan-to-value at which it liquidates

   A market decides the price at which you are liquidated. That is a term of the loan, not a detail, and it can differ between two venues quoting the same rate.

4. Supply APY

   What the other side earns. Relevant when you are choosing where to put money rather than where to take it from.

The trap that costs more than the spread#

Here is the thing worth the whole post, because it is a genuine mistake people make and the numbers look fine while they make it.

Compare that with Aave v3, where a supplied asset both backs your borrow and earns. The two venues use the same word for structurally different things, and the gap is larger than any rate spread you were optimising.

Why “cheapest” is the wrong optimisation anyway#

A borrow is not a product you buy once. It is a position with terms.

On headroom specifically: aim well below the ceiling. Collateral bought by an earlier step arrives worth slightly less than was spent, so a plan targeting the maximum loan-to-value starts out tighter than intended. Half is a reasonable default.

Moving the loan, when it is worth it#

If the spread does justify a move, the mechanics are the reason this is worth automating.

By hand, refinancing a loan is a deadlock: to repay the old loan you need your collateral back, and to get your collateral back you have to repay the loan. Your only way through is to find the entire loan amount somewhere else first.

The money that breaks the deadlock is borrowed and repaid inside the same transaction. Either the whole thing happens or none of it does, and your account never holds the borrowed amount.

How to actually get the answer#

One call, not one per venue.

What comes back is every market’s supply APY, borrow APY, utilisation and liquidation loan-to-value, across Aave v3, Aave v4, Morpho and Revert vaults on all three networks. Omit the filters entirely and you get all of it.

It costs nothing. Of fifty-two tools, five spend money and this is not one of them.

The answer to the original question#

There isn’t one number, and anybody who gives you one is quoting a venue rather than answering.

What there is: a spread worth more than a point between venues most people treat as the same, a utilisation figure that tells you whether today’s rate survives contact with next week, and one venue where the collateral you supply earns exactly nothing.

Ask for all of it at once, then borrow. The asking is free; only the borrowing costs.