Tether Earned $1.5 Billion and Cut Its Safety Margin in Half
Profit and safety are not the same number. Tether's second quarter is the cleanest demonstration of that anyone has published this year.
The largest stablecoin issuer in the world reported roughly $1.5 billion in net operating profit for the three months ending June 30. Over the same period, its excess reserves — the cushion that sits above the dollar-for-dollar backing of USDT — fell to $4.11 billion from a record $8.23 billion. Tether earned at a rate most regional banks would envy and finished the quarter with half the margin for error it began with.
That isn't a contradiction. It's a choice. And it deserves considerably more attention than the profit headline got on Friday.
The cushion is the only part of a stablecoin that can absorb a bad day
Every dollar of USDT is supposed to be matched by a dollar of assets. Fine. But assets move. Treasury bills reprice, secured loans go bad, gold swings, bitcoin swings harder. When the asset side dips and the liability side doesn't, something has to sit in between. That something is the excess reserve.
Tether's BDO-prepared attestation put total assets at $187.75 billion against liabilities of $183.64 billion on June 30. Do the division. The buffer is 2.24 percent of what Tether owes its holders.
Set that next to a US bank. A depository institution needs a Tier 1 leverage ratio of at least 5 percent to be called well capitalized by its regulator — and banks hold assets far more varied and far more illiquid than short-dated government paper. Tether's reserve mix is genuinely more conservative than a bank's loan book. Its buffer is also less than half the bank threshold.
Both things are true. Neither cancels the other.
The quarter was profitable and expensive at the same time
Here's the part that gets lost. A $4 billion drop in the cushion alongside $1.5 billion of profit means something on the order of $5.5 billion left the buffer through channels that aren't operating losses.
Tether has been open that it moves capital out of the reserve stack and into its own balance sheet — investments in Rumble, in agricultural commodities trading, in mining infrastructure, in a sprawling portfolio that has nothing to do with keeping a token pegged. Those investments may be brilliant. They are not liquid at par on a Tuesday morning when redemptions spike.
The company also expanded its gold holdings past 146 tons in the quarter. Gold is a fine store of value and a terrible redemption mechanism. Nobody wires bullion to a market maker at 3 a.m.
So the buffer didn't evaporate. It converted — from the thing that absorbs shocks into the thing that generates returns. That trade has an obvious upside and a non-obvious cost, and the cost only shows up on the worst day.
The pie is shrinking, which makes the ratio harder to fix
If the stablecoin market were still compounding, none of this would matter much. Growth papers over thin cushions; new issuance brings new reserve assets and the buffer rebuilds itself.
It isn't compounding. Stablecoin market capitalization fell 1.6 percent to $305.1 billion in the second quarter, the first contraction since the third quarter of 2023. It has kept sliding since, sitting near $300.8 billion as of yesterday. Measured from the May 20 peak around $321 billion, this is the third-largest drawdown in the sector's history — behind only the 2022 collapse.
What makes this one strange is how quiet it is. USDT has traded in a band between roughly $0.9988 and $0.9992 through the whole decline. USDC has mostly held above $0.9997. There is no run. There is no panic. There is just money walking out the door in an orderly line, mostly because Treasury yields came down and holding a dollar token stopped paying what it used to.
Circle's numbers tell the same story from the other side. Its reserve return rate dropped 66 basis points to 3.5 percent, revenue came in at $701 million against a $717 million consensus, and USDC circulation ended the period down 4.8 percent at $73.3 billion. Circle's answer has been to chase transaction share rather than float — its on-chain volume rose 151 percent to $14.8 trillion, and USDC hit nearly 70 percent of stablecoin transfer volume in June.
Two issuers, two strategies. Circle is betting the future is in payment throughput. Tether is betting the future is in the balance sheet it built with the float. USDT's market share climbed to about 60 percent while all this happened, so the market is not exactly punishing the second bet.
Attestations are not audits, and everyone knows it
Tether has never produced a full financial statement audit from a Big Four firm. It produces quarterly attestations — a narrower engagement, a point-in-time snapshot, a different standard of assurance. BDO signs them. Critics have hammered this for years and Tether has spent those years growing anyway.
The GENIUS Act, signed last year, tightens what a payment stablecoin issuer operating in the United States must hold and disclose. Tether's response has been to run a separate US-compliant token while USDT itself remains an offshore instrument for offshore users. That's a legally coherent structure. It also means the token holding 60 percent of the market is the one furthest from the new rulebook.
Nobody should read a thin buffer as a prediction. Tether has survived every mortality forecast written about it since 2017, including one from a sitting US regulator. Its reserve composition today is dramatically better than the commercial-paper stack that drew the New York Attorney General's attention in 2021. Progress happened.
But the direction of travel this quarter was backward, and the company chose it.
The number your app shows you is the least informative number available
Open a wallet. USDT shows $1.00. It showed $1.00 in April when the cushion was $8.23 billion, and it showed $1.00 in June when it was $4.11 billion. The interface is identical. The underlying risk changed by half.
That gap between what a holder sees and what a holder holds is the actual problem with stablecoins in 2026 — not depegs, which are rare and brief, but the total absence of reserve quality in the surfaces where people make decisions. Portfolio tools like The Crypto App can show you what you own down to the wei, across chains, in real time. None of them can show you how much of a shock the thing you own is built to absorb, because issuers publish that quarterly, in PDF, ninety days after the fact.
Traditional finance solved this crudely with credit ratings. Crypto hasn't solved it at all.
What a stablecoin actually competes on now
For five years the pitch was distribution. Get on more chains, more exchanges, more remittance corridors, more merchant rails. That race is largely settled — competition has moved from issuing tokens to owning distribution, and the winners in distribution are increasingly banks and payment processors rather than crypto-native issuers.
Which leaves reserve quality as the last differentiator nobody markets on. There is no issuer running campaigns about their cushion ratio. There is no dashboard ranking stablecoins by loss absorption. The one variable that determines whether your dollar survives a genuinely bad week is the one variable the industry has agreed not to compete on.
That agreement holds right up until it doesn't. And when it breaks, it won't break gently — it will break the way it always breaks, with everyone discovering simultaneously that they were looking at the wrong number the entire time.
The next stablecoin winner may not be the one with the most integrations or the fattest yield. It may simply be the first one willing to publish its buffer daily and let the market watch.