Where the yield actually comes from
1 September 2026 · written for Treasurers, allocators and depositors considering a USDC position
Yield that cannot be explained should not be trusted. This is the plain version: who is borrowing your USDC, what they had to put up to borrow it, what makes the rate move, and what would have to happen for it to stop.
Most people evaluating a yield product start with how much. It is the wrong first question, and it is the one every bad product answers most confidently.
The better first question is where it comes from. A rate is an output. If you cannot trace it back to somebody paying it, and say why they are willing to, then you are not looking at yield — you are looking at a number on a page, and the difference only becomes apparent later.
This is the traceable version, written for someone who does not work in crypto.
Your USDC earns because it is lent, and what a borrower pays to borrow it is the yield. That is the whole of it. There is no trading, no strategy, no counterparty cleverness in the middle taking a view on markets.
The natural next question is who borrows USDC, and why they would pay to.
Mostly people who hold an asset they do not want to sell. Someone holding bitcoin who needs dollars for a few months can either sell it — a decision that is permanent, and taxable in most places — or borrow dollars against it and keep the position. Trading desks borrow to fund inventory. Institutions borrow against treasuries or tokenised real-world assets. In every case borrowing is cheaper than selling something they expect to want later.
That demand is not sentiment about crypto. It exists because selling and borrowing are different decisions with different consequences, which is as true in traditional finance as anywhere else.
The obvious worry is what happens when a borrower does not repay. The answer is that they have already posted more than they borrowed.
To borrow, they lock collateral in the contract first, worth meaningfully more than the loan. If the value of that collateral falls toward the size of the loan, the position is closed automatically and the collateral is sold to repay the lenders. That process is mechanical. It does not require the borrower to cooperate, and it does not wait for anyone to make a decision.
This is the part with no equivalent in a bank deposit. A bank lends your money to someone whose ability to repay is a judgement about their future income. Here, the repayment is secured by an asset already held by the contract, and the margin between the two is set in advance.
It is not risk-free. Nothing is. But the failure mode is specific and worth naming: collateral that falls faster than it can be sold, in a market too thin at that moment to absorb it. That is the scenario the whole design is built to survive, and it is the one to ask about.
The rate is not set by us and not set by anyone. It is a function of how much is deposited against how much is borrowed.
When borrowing demand rises against available supply, the rate rises, which draws more deposits and cools it. When deposits pile up unborrowed, the rate falls, which pushes lenders elsewhere. It is a thermostat, and it is running continuously.
Two consequences follow, and both matter more than any particular number.
The rate is variable, always, and never a promise. Anything presenting a fixed return on a variable market is either hedging that risk somewhere you cannot see, or hoping.
And a rate that is unusually high relative to everything else is information rather than opportunity. It means the market is demanding more to hold that particular risk. The right response is to find out what the market knows, not to assume you found something it missed.
Three things, and it is worth knowing them by name.
Collateral could fall too fast to liquidate cleanly, leaving a shortfall that lenders absorb. This is the primary risk and it is not hypothetical — it has happened elsewhere in lending markets, and the honest question is not whether it is possible but how much margin sits in front of it.
The price feed the contract reads could be wrong. Liquidation depends on knowing what the collateral is worth; a feed that is stale or manipulable is a liquidation that fires too late or for the wrong amount.
Or everyone could try to leave at once. Deposits are lent out, so they cannot all be withdrawn instantly any more than a bank could return every deposit on the same afternoon. When utilisation is high, withdrawals wait for borrowers to repay or for new deposits to arrive. The rate rising is what resolves it — which is the thermostat working, not failing, though it does not feel that way while you are waiting.
None of these are exotic. They are the risks of secured lending, which is a business with a long history and well-understood ways of going wrong.
Somebody chooses which collateral is acceptable, how much margin to require, and which markets to enter at all. Those choices are what separate a conservative position from an aggressive one, and they matter more to your outcome than the rate on the day you deposit.
Those decisions belong to independent curators — professional risk managers who publish their methodology and are accountable for it. They set the parameters. They are not us, they do not work for us, and they can change their parameters without reference to us.
That independence is deliberate. A product where the same party sets the risk, holds the money and reports the results has no check in it anywhere.
It is not a bank deposit. There is no insurance, no guarantee, no institution standing behind it, and no rate that holds.
What replaces those assurances is visibility. Every market, every parameter, every position and every rate is readable on-chain, continuously, by anyone — including you, including now, and without asking permission.
That is a genuinely different bargain from the one a bank offers, and it is worth being clear-eyed about which parts of it you are getting and which parts you are giving up.
Everything above is verifiable, and none of it requires taking our word.
Ask any yield product where the yield comes from and expect a specific answer: who borrows, what they post, what happens when they cannot repay, and who sets those terms. Ask what the rate does when demand falls. Ask what happens when everyone withdraws at once.
A product that cannot answer those plainly is not necessarily dishonest. But it is asking for trust it has not shown you the basis for, and you are entitled to more than that before you move money.
dyield tokens are yield-bearing vault receipt tokens, not stablecoins, not bank deposits, not insured.