What the Platform Does With the Stablecoins You Deposit

Once your USDT goes into an Earn product, it doesn't sit in a drawer marked "your account". It gets used, it generates a return, and part of that comes back to you. Working out where it went is far more useful than staring at the APY.

Path one: lent to people wanting leverage

The mainstream path and the easiest to understand. There are people on the exchange who want to trade with leverage and don't have enough capital, so they borrow USDT. They pay interest, and the platform passes some of it to depositors.

Risk on this path is relatively controllable: borrowers post collateral, and the collateral ratio is normally well above the loan.

When the price moves against them, the system first calls for more margin and then force-liquidates, and the purpose of that liquidation is precisely to make sure the lent money comes back.

Its problem isn't bad debt, it's volatility: more borrowers means higher rates; a quiet market with no borrowers means rates fall. Your return follows that supply and demand, which is the root reason the Flexible APY changes daily, expanded in how borrowing demand drives the rate.

What happens in extreme markets

When the price gaps hard, forced liquidation may not execute at the intended level, producing a shortfall. Mature platforms absorb those losses with a risk reserve, which is why large platforms usually disclose that they hold an insurance fund. To judge whether that fund is meaningful, look at two numbers: the size of the fund, and the platform's total open positions. When those differ by orders of magnitude, the fund covers ordinary volatility and not the systemic event.

Path two: market making and liquidity

The second path puts the money to work providing liquidity: quoting both bids and offers on a pair and earning the spread, or supplying certain liquidity pools and earning a share of fees.

Returns here come from trading activity, the more volume, the more earned. The risk comes from prices moving fast in one direction: a market maker keeps getting filled on the way down and accumulates an asset that is falling. Traditional finance calls this inventory risk; in crypto, with bigger swings, it shows up more sharply.

For an ordinary depositor, the key question on this path isn't whether it's dangerous, it's whether you can even find out that your money took it. Most platforms' product descriptions don't go to that level of detail.

Path three: the platform's own arrangements

The third is the vaguest and the one that needs the most caution: the platform uses depositor funds for its other businesses: proprietary investing, institutional lending, participating in certain projects, or simply filling a hole somewhere else.

Most of the big failures in crypto over the past few years relate to this layer. On the surface, a "high-yield Earn product"; underneath, depositor money deployed somewhere opaque that couldn't be repaid when the market turned. Before it blew up, the user interface showed no difference at all — same subscribe button, same APY, same credited returns.

That isn't to say every platform does this. It's to say: you can't judge this layer from a product page. You can only judge it from the platform's overall transparency.

What "your coins" actually are inside a platform

There's a question more basic than where the money goes, and most people have never asked it: when you put coins into an Earn product, what does the platform's ledger record?

The answer is an accounting entry, not an on-chain asset that belongs to you. Your coins and everyone else's are commingled in the platform's pool, and the platform tracks internally how much each person has. That's how centralised platforms have always worked, and traditional finance is the same. There's nothing shady about it in itself.

But it has a few direct consequences worth knowing.

You and the platform are in a creditor relationship

Strictly, what you hold is a record of how many coins the platform owes you. While the platform operates normally, that record and the actual coins are indistinguishable. If the platform fails, you are a creditor and you join an insolvency process and wait. Historically those processes take years to complete, and don't necessarily return the full amount.

You can't find "your" coins on-chain

People go to a block explorer looking for their coins, can't find them, and panic. Not finding them is normal; your coins are in the platform's aggregated addresses, mixed with everyone else's. The only way to have an asset on-chain that's yours is to withdraw to a self-custody wallet, and that carries a different set of risks (lose the private key and nobody can help you).

What "proof of reserves" actually proves

It proves the platform's total assets cover its total liabilities to users: that is, nothing has been misappropriated to the point of insolvency. It doesn't prove what any individual asset is being used for, and it isn't live: what's published is a snapshot at a point in time, and the edition you're reading may be one or two months old. So it answers a narrow question: on that day, the books balanced.

How much disclosure is enough

Disclosure varies a great deal between platforms. Reading it in the following layers tells you where a given platform sits.

1. A product page only

   An APY and a term, with nothing at all about what the money is used for. The bottom rung.

2. A product description and risk warnings

   The description mentions that funds go to lending or liquidity and lists the main risks. Already a lot better than the first rung.

3. Proof of reserves

   User asset reserves published on a schedule so outsiders can check that the platform holds enough. The verification method and coverage differ between platforms, so it's worth looking at what exactly was checked.

4. Independent audit or third-party verification

   An external firm checks and issues a report. The higher rung in this industry today, but note the scope of the audit and the date it covers.

Binance publishes pages relating to proof of reserves; the current explanation and latest data are whatever the official help centre shows at the time. We don't restate figures here; that data updates on a schedule and would go stale the moment it's written down.

Where it differs from a bank

Many people are reasoning about stablecoin Earn with bank-deposit intuitions, and that analogy fails at several key points. Laying the differences out sharpens your sense of the risk.

Whether there's a backstop

Bank deposits in most countries carry deposit insurance, so up to a limit you get your money back even if the bank fails (for the US version see the FDIC's deposit insurance explanation; schemes differ by country).

There's no equivalent on the stablecoin Earn side. A platform's own risk reserve is a buffer whose size and rules for use the platform sets itself.

Whether there's a lender of last resort

A bank in a liquidity crisis can borrow from the central bank. A crypto platform has no such channel and has to meet concentrated redemptions out of its own liquidity. So a run of the same size does its damage faster here.

Different regulatory intensity

A bank's capital adequacy, liquidity coverage and asset allocation are all subject to continuous supervisory review.

On the crypto side it varies by jurisdiction — strict in some places, almost absent in others. So when you see the word "Earn", it's worth checking who regulates that platform where you live and whether it holds a local licence. The same word can sit on wildly different constraints.

Different stability in the source of the return

A bank's net interest margin is relatively stable because borrowing demand is relatively stable; crypto borrowing demand follows the market, which is why the Flexible APY moves daily. That difference has a direct corollary: the days with the highest APY are usually the days the market is most excited.

Why this is worth being pedantic about

If this were only a theoretical risk it wouldn't deserve this much space. The problem is that it has genuinely happened, more than once.

Around 2022, a batch of products marketed as "high-yield savings" appeared in crypto, and in the user interface they looked no different from legitimate products: enter an amount, tap confirm, watch the return tick up daily. Underneath, depositor funds were deployed into illiquid, maturity-mismatched or entirely opaque places. When the market turned, those firms became unable to pay out one after another, and users' money was frozen inside.

Reviewing them afterwards shows a common thread: before failing, all of those products paid clearly above their peers, and not one of them could explain where the return came from. The only difference a user could see when choosing was the higher number, which is precisely the thing that should never be the basis for choosing.

Another kind of risk comes not from bad intent but from structure: taking short-term redeemable deposits and putting them into long-term assets you can't exit. Fine in normal times; the moment a lot of users redeem at once, the liquidity snaps. The traditional banking system covers that with deposit insurance and a lender of last resort, and crypto platforms have neither.

How an ordinary person can judge

A few crude methods that work without doing due diligence.

Check whether the APY is persistently abnormal. A short burst of high interest may be a campaign; sustained high interest always has a cause. The typical patterns behind abnormal rates are in the patterns behind abnormally high yields. If it sits far above comparable products over time, ask why first.

Check whether the description says what the money is for. A platform willing to spell it out is at least being straight with you on that point. One that talks only about returns leaves you without the material to judge at all.

Check whether it tells you uncomfortable things. A product page full of risk warnings, redemption conditions and how extreme situations are handled is unpleasant to read, and that's exactly what telling you the real situation looks like. A page that's nothing but upside is the one to be wary of.

Don't keep everything in one place. Clichéd, but it's the only method that doesn't depend on your judgement being right. In practice, set yourself a ceiling: no more than some share of your total position on any single platform, and don't break it because one of them "looks especially solid".

Four questions you can go and answer yourself

With the disclosure layers covered, here's the concrete action: spend twenty minutes looking up the four questions below. It isn't due diligence, but it beats doing nothing by a long way.

None of the four requires expertise — only that you're willing to click.

One: how does the platform describe what this product's funds are used for

Search the product page and the help centre for anything about how the money is used. If you find it, read it, and note whether it uses definite or vague language. Finding nothing at all is itself an answer.

Two: where are the reserves published, and how often are they updated

Find that page, check when the latest edition is, then go back two or three editions and see whether the interval is stable. Monthly publication suddenly going quiet for three months is a change worth more attention than any single edition's numbers.

Three: what do the redemption terms say about extreme situations

The user agreement usually has a section on this. Don't get stuck on legal language; just look at which rights it reserves: can it suspend redemptions, can it delay them, does it commit to notifying you. If you can't find the section, search for "suspend", "delay" and "force majeure"; they're usually towards the back.

Four: who regulates this platform where you live

Whether there's a local licence and which body supervises it decides whether you have anywhere to complain if something goes wrong. This information is usually in the site footer or the About page, and platforms differ in how much they disclose.

Having answered all four, you probably still can't conclude "safe" or "unsafe"; that was never a judgement an ordinary user could make. What you get is four specific records: how they describe fund usage, how often reserves are published, which rights the agreement reserves, and who regulates them where you are. Write those four down, and you'll notice immediately if any of them changes later.

Signals worth watching

Platforms rarely fail without warning. Ordinary users can't get inside information, but some public signals are observable.

Withdrawals starting to slow down or acquiring extra conditions. The most direct signal. A platform operating normally has a stable withdrawal process; requests for extra documents, longer reviews, or new limits (with a vague explanation) are worth taking seriously.

A sudden sharp increase in deposit rates. Raising rates on its own while peers' rates haven't moved usually means it needs funding more badly. This isn't the same as a campaign boost: campaigns have a defined deadline and cap, while this kind is broad and sustained.

Key people leaving in numbers, or going quiet. Team changes visible in public information, social accounts going stale, no response to questions; individually these mean little; appearing together, they're worth noticing.

Reserve reports delayed or their scope changing. Monthly publication becoming irregular, or the report's coverage quietly shrinking, are both things to ask about.

What should you do when you see these? Not immediately pull everything out and broadcast it everywhere: that can both damage a good platform unfairly and push you into a worse decision while panicking. The reasonable response is to reduce exposure, stop adding, and keep watching. Bring the position down to a level where you could accept even the worst outcome, then carry on observing.

On whether the stablecoin itself can fail, we went through the historical depegs separately in when the stablecoin itself goes wrong. The product-structure comparison is in the product structure layer.