The number that should stop you is not $187.75 billion. It is $4.11 billion.
On June 30, Tether's reserve disclosure listed $187.75 billion in assets against $183.64 billion in liabilities. The difference โ $4.11 billion โ is the entire equity cushion standing between USDT and a redemption queue. That is 2.2% of the float. A U.S. money market fund operating under Rule 2a-7 is not permitted to let weekly liquid assets slip below 30% of total assets without triggering a formal review. Tether's excess reserve, as a share of what it owes, is roughly one-fifteenth of that level. USDT is not a fund. Most of the people holding it, however, behave as if it were. That gap between what the instrument is and what its holders believe it to be is the first anomaly worth naming, and it is the one that survives every cycle. The ledger never lies, only the narrative does. So let me read the ledger.
Context
Tether is not a protocol in the sense that most of this industry uses the word. It has no governance token, no validator set, no emissions schedule, and no on-chain voting. It is a centralized issuer that mints a liability on many chains and holds an off-chain asset pool against it. The two-layer structure matters because it is where the confusion lives: the token is on-chain and verifiable, the backing is off-chain and reported. You can audit the first. You must trust someone for the second. Trust is a variable I do not solve for โ I price it.
Let me be precise about the reporting layer, because it is constantly misdescribed. There are two distinct documents, and they are not interchangeable.
A reserve attestation is a point-in-time statement from an accounting firm that the figures management presented are consistent with the records reviewed. It is a snapshot. It does not test internal controls, does not trace every asset to a custodian, and does not opine on whether the reporting framework is sound. In August, KPMG US issued an unqualified โ clean โ opinion on Tether's 2025 financial statements. That is a genuine milestone. An unqualified opinion is the highest tier of assurance an auditor can give, and it is a meaningful upgrade from the attestation era that critics spent years attacking.
But the two documents answer different questions, and the headline has blurred them. A clean audit of the financial statements is not a line-by-line verification that every dollar of reserve exists at every moment. It is not a government guarantee. As the source material behind this analysis puts it, none of these documents makes Washington treat USDT as an obligation of the United States. That sentence is the spine of everything that follows.
One more framing point before the data. Stablecoins of this type generally do not fail the Howey test for securities status โ there is no expectation of profit, no common enterprise driven by the issuer's managerial effort, and no upside promise to holders. That regulatory floor is real. But clearing the securities question does not clear the money-transmission, reserve-disclosure, and sanctions-compliance questions, which is where the actual oversight pressure sits. The instrument is not a security. It is also not a bank deposit, and it is not a Treasury. The category matters because each category carries a different promise, and USDT carries a narrower one than most holders assume.
Core
Start with the asset stack, because what backs USDT is now a portfolio, not a bank account.
As of June 30, the composition breaks down like this: $114.96 billion in directly held U.S. short-term Treasuries, or 61.2% of reserves; $18.63 billion in overnight reverse repurchase agreements, 9.9%; $18.84 billion in precious metals, 10.0%; $5.8 billion in Bitcoin, 3.1%; and $13.45 billion in secured loans, 7.2%.
Read that list twice. The story most people carry in their heads is that USDT is backed by Treasuries. The data says 61.2% is Treasuries. The remaining 38.8% is a mix of highly liquid instruments, volatile instruments, and credit exposures. The overnight repo is effectively cash-equivalent and I treat it as such. The precious metals and Bitcoin are mark-to-market positions that move with markets Tether does not control. The secured loans are the line item that should keep you up at night, and I will come back to them.
I want to pause on the source's deliberate separation of the $18.63 billion overnight reverse repo from the $114.96 billion Treasury line. On a conservative reading, both are top-tier liquidity, and collapsing them is defensible. But the distinction exists for a reason: reverse repo is a short-dated money-market instrument whose yield and availability depend on counterparties and the funding market, while a directly held T-bill is a claim on the U.S. Treasury itself. The gap between the two is small in normal conditions and non-zero in stressed ones. When a reserve report insists on splitting them, it is telling you that the quality of the reserve has a discretionary margin โ and discretionary margins are where narratives expand to fill space the numbers leave open.
Alpha hides in the variance, not the volume. Everyone quotes the $187.75 billion total. Almost no one decomposes it. So let me decompose the duration and liquidity profile, because that is where the real risk sits.
Treasury bills under one year carry minimal duration risk โ if rates move, the mark moves a little, and principal is returned at par quickly. That is the safe bucket. The overnight repo is the safest bucket; it resets daily. Together, Treasuries plus overnight repo equal $133.59 billion, or roughly 71% of reserves. That 71% is the genuinely liquid, low-volatility core. It is why a redemption run, in the absence of some exogenous crisis, would not instantly break the peg.
Now the other 29%. Precious metals and Bitcoin together are $24.64 billion. These are not bad assets. They are volatile assets. In a sharp risk-off move โ the kind this bear market keeps threatening โ a 30% drawdown in that basket erases roughly $7.4 billion of reserve value. Tether's entire excess reserve is $4.11 billion. Do the arithmetic yourself: a simultaneous hit to metals and Bitcoin, plus any impairment on the loan book, is larger than the cushion. The buffer can be wiped by a market move, not by fraud.
That is the finding I want to land. The primary risk in USDT is not that the reserves do not exist. It is that the reserves are not the conservative, Treasury-only instrument the narrative implies, and the equity buffer protecting them is thin. In a bull market, a 2.2% buffer is a rounding error nobody discusses. In a bear market, it is the whole conversation, because survival is the only return that compounds.
I have done this kind of work before. In 2017, while running quantitative due diligence at a mid-sized crypto fund in Denver, I audited 45 whitepapers and tokenomics models and flagged the structural absurdity of emission schedules that could not survive contact with their own roadmaps. That exercise taught me a habit I have never dropped: reconcile the supply schedule against the balance sheet, every time. The same habit applies here. The supply is $183.64 billion of liabilities. The balance sheet is $187.75 billion of assets. The reconciliation gap is $4.11 billion. That is the whole margin for error, and the balance sheet carries volatility.
Let me turn to the profit mechanism, because it explains why Tether will keep doing this regardless of sentiment.
In the second quarter of 2025, Tether reported roughly $1.5 billion in net operating profit, driven almost entirely by interest on Treasuries and reverse repo. Annualized, that is a business throwing off something in the neighborhood of $6 billion. And here is the structural point that gets buried: USDT holders have no contractual right to any of that income. You deposit a dollar, you receive a dollar-denominated token, you earn nothing. Tether keeps the spread between the yield on the reserve and the zero it pays to holders.
Compare that to a money market fund. An MMF holds short-duration, high-quality assets and passes the yield through to investors as dividends. Tether runs the same asset strategy and passes through nothing. It is, functionally, a non-yielding money market fund with a crypto distribution network bolted on top.
I want to be careful about what this is and is not. This is not a Ponzi structure. A Ponzi pays early participants with new entrants' money and has no real asset base. Tether has a real asset base and real, risk-free-ish income. The revenue is genuine. But not a Ponzi is a low bar, and clearing it does not make the arrangement fair. The value capture is asymmetric and it favors the issuer by design. That asymmetry is a structural feature, not a temporary condition, and it creates a specific forward risk: the moment a competitor offers a yield-bearing dollar that is credibly safe, the marginal USDT holder has a reason to migrate. I flagged a version of this in 2020, when I backtested yield strategies across Aave and Compound and found that simple rebalancing beat complex leverage by 15% in volatility. The lesson then was that capital flows to the cleanest risk-adjusted return, not the loudest one. The same current runs under stablecoins.
The engine, though, is rate-sensitive, and this is the part the bulls underweight. Tether's revenue is essentially the risk-free rate multiplied by the size of the liquid reserve. If the Fed cuts, the yield on $114.96 billion of T-bills and $18.63 billion of repo compresses together. A 100 basis point cut on a roughly $133 billion liquid book is on the order of $1.3 billion of annualized revenue gone. The reserves do not shrink, but the capacity to grow the excess buffer โ the very cushion I have been worried about โ depends on that revenue. Rate cuts do not make USDT unsafe. They make it slower to defend itself.
Now the black box.
Of the $187.75 billion, $13.45 billion sits in secured loans. The counterparties are not disclosed. The collateral is not disclosed. The loan-to-value ratios are not disclosed. This is the single largest information gap in the entire structure, and it is larger than the excess reserve by a factor of three. If any meaningful share of that book sours โ a borrower fails, collateral turns out to be illiquid or overvalued โ the loss lands directly on the $4.11 billion cushion. I cannot price that risk with the data available. What I can do is refuse to pretend it is not there. An undisclosed loan book three times the size of your equity buffer is not a footnote. It is the balance sheet.
There is a second-order point here. Tether holds $5.8 billion in Bitcoin. A stablecoin issuer holding Bitcoin in reserve is making a directional bet with depositor backing. That may be a deliberate strategy to diversify and capture upside, but it is philosophically in tension with the conservative mandate that a dollar token claims. A dollar instrument that carries Bitcoin exposure is a dollar instrument with an equity-like tail.
Which brings me to the shift nobody fully priced: the political relationship.
In 2021, the CFTC fined Tether $41 million for misleading statements about its reserve backing, tied to claims made between 2016 and 2019 that USDT was fully supported by fiat in bank accounts when it was not. In October 2022, Tether cleared commercial paper from its reserves and replaced it with short-term Treasuries. And in 2025, it obtained a clean audit opinion. The arc runs from penalized misrepresenter to audited counterparty.
Meanwhile, per Bloomberg reporting, the Trump administration is weighing an overseas stablecoin initiative involving the Treasury, the State Department, and the DFC. The reporting has not confirmed any deal with Tether, and no plan has been formally announced. Read that last sentence again, because it is doing a lot of work. What exists is a reported idea, not a signed contract. The narrative has already run ahead of the evidence.
Still, the direction is legible. A firm operating largely outside the conventional U.S. banking system has become a major buyer of U.S. debt. Washington gets demand for its paper without running a retail operation. Tether gets the interest. Users get a transportable dollar balance they can hold in places where opening a U.S. bank account is impractical. Three parties, three interests, one instrument. That is a real alignment. It is also a dependency. Tether's standing now rests substantially on the current political disposition of a single jurisdiction. The same Washington that can elevate it can, on a different day, re-prioritize the CFTC fine, push for reserve mandates, or license a competitor. In 2022, after Terra collapsed, I spent six weeks reconstructing the death spiral from reserve proofs and redemption delays, and the lesson that stuck was mechanical: what breaks a system is rarely a single shock; it is a dependency you assumed was stable. Here, the dependency is political goodwill.
Contrarian
Here is where I part company with both camps.
The bulls point to the clean audit and the Treasury holdings and conclude USDT is now safe. The bears point to the 2021 fine and the Ponzi line. Both are reading the wrong variable. The audit is a verification of the financial statements, not a real-time proof of reserves, and the reserves are a mixed portfolio, not a Treasury-only vault. The bears are also wrong, because the income is real and the assets exist โ this is not UST, and treating it as such is a category error that has already cost people money on the wrong side of the trade.
Correlation is not causation, and proximity is not guarantee. The fact that Tether holds U.S. Treasuries does not make USDT a U.S. obligation. Buying a Treasury is not the same as being a Treasury. This is the deepest logical flaw in the government-backed story, and it is exactly the confusion the source material flags when it notes that none of these documents transform USDT into a sovereign liability. Tether is a private issuer whose liabilities are Tether's, not America's. The narrative of sovereign backing is doing sentiment work the legal structure does not support.
There is a second contrarian read. If the political embrace is real, it is shared. A policy shift that legitimizes dollar stablecoins lifts the entire compliant sector โ including USDC, which competes directly and runs a more conservative, more transparent reserve. Institutional capital that enters on the back of a stablecoins-are-now-legitimate headline does not necessarily enter through USDT. It may well enter through the product with the cleaner balance sheet. The beneficiary of the narrative and the beneficiary of the flows are not guaranteed to be the same token.
And the timeline risk cuts both ways. The overseas stablecoin plan is reported, not enacted. If it stalls, the sentiment premium it created deflates. If it proceeds with an official or joint-venture issuer, that competitor could dilute the very market share Tether currently claims. Remember that the 60%-plus share figure is self-reported. I have not seen an independent, cross-verified measurement in the material I am working from, and self-reported market share deserves the same skepticism I applied to self-reported reserves in 2017.
Takeaway
Watch the reserve report, not the headline. The signal that matters over the next cycle is the composition of the excess reserve and the disclosure quality of the $13.45 billion secured loan book. If the buffer thickens and the loan counterparties become visible, the structure is de-risking and the faith is earned. If the buffer stays near 2.2% while the volatile and illiquid share of reserves grows, then the market is holding a dollar token whose safety margin depends on markets staying calm โ and this is a bear market, where calm is the scarcest asset there is. In 2024, when I tracked ETF inflows against exchange outflows and found long-term holder accumulation rising 12% while reserves fell, the takeaway was the same as it is now: the flows tell you what the balance sheet will look like before the price does. Due diligence is the only hedge against chaos. Read the ledger. It is the one document in this entire structure that does not get to lie.