Five cents of every dollar Washington collects in taxes now goes to interest. Not to defense. Not to roads. To the coupon. Net interest crossed $1.1 trillion, up 11% year over year, making it the third-largest federal line item behind Social Security and Medicare.
Here's the part that stopped me. The average rate on the entire federal debt rose just 0.06 percentage points, from 3.415% to 3.475%. Six basis points. The interest bill jumped double digits anyway.
You don't get an 11% increase from six basis points of rate. You get it from volume.
That gap โ flat rates, exploding cost โ is the entire trade. It's also the number every crypto desk should be watching, because the same mechanics inflating Treasury coupons are inflating the reserves backing every dollar of USDT in circulation.
Structure before conclusion.
The fiscal-year 2026 deficit runs near $1.993 trillion, up 12%, close to 6% of GDP. Read that twice โ 6% during an expansion. Structural deficits don't shrink in a recession. They widen mechanically.
The 10-year sits at 5.35%. A $39 billion auction cleared at 5.3%. One-third of all marketable Treasury debt matures inside twelve months and must be rolled at those yields.
Meanwhile the Fed runs QT. It is not buying. Supply rises while the largest historical buyer walks away. Two forces pushing the same direction on price. That's not a cycle. That's a term-premium regime.
One number the source never splits: the 10-year yield is three things stacked โ real rate, inflation expectation, term premium. If the climb is inflation, the Fed has a job. If it's term premium, the Fed has a problem. The weak payrolls say inflation is cooling. The auction data says the premium is widening. That points to the second case, and the second case is the one monetary policy cannot fix.
And the political layer makes it worse. Before the midterms, neither party lists deficit reduction as a priority. Meanwhile a proposed $5,000 per-adult check carries a price tag above $1 trillion, financed โ if it happens โ by more borrowing. Cut the deficit and you lose votes. Expand transfers and you win them. That asymmetry is the real budget process, and it points one direction: more issuance. Policy promises are now a leading indicator of supply.
I'll flag my own skepticism. The source's timeline doesn't reconcile โ a 5.35% 10-year sits above anything in the verifiable record, and the fiscal labels collide with the yield history. I treat the absolute levels as low-confidence. The structural relationships hold regardless. Those are what I trade.
Here's the empirical tell I trust: interest cost is rising faster than the average coupon. That is a rollover signature, not a rate signature. The pain is scheduled, not speculative.
The refinancing wall is arithmetic, not opinion. If a third of the debt rolls within a year, and it rolls from an average coupon near 3.4% into a market clearing above 5%, the blended rate climbs whether or not the Fed ever moves. The policy rate sets the front. The auction calendar sets the back. They are now two different clocks.
This is fiscal dominance in its mechanical form. When the long end is priced by supply, the policy rate stops being the anchor. You can cut the front and watch the back rise. That combination has a name in the literature and almost no name in retail flow.
Now the crypto transmission. This is what I actually trade.
Tether's reserve book is a short-duration Treasury fund wearing a stablecoin costume. The overwhelming majority of its backing is T-bills. When the front end yields 5%, Tether earns roughly $5 billion a year on a ~$100 billion book before it pays a cent to holders. That's the business model. It's also why the reserve question never resolves. An independent audit would put a hard number on the float, and the float is the margin. ZK proofs don't fix this โ a cryptographic attestation is only as honest as its inputs, and the input here is a custodian's word. Code is law, but gas fees are the reality; a reserve is a promise, and the market is the enforcement.
Follow the yield. When T-bills pay 5%, the delta-neutral stablecoin trade โ short the perpetual future, long the bill โ prints a spread that retail chases as "passive income." The basis is not free money. It is a levered bet that the funding rate stays positive and the collateral holds value. A rising term premium stress-tests both legs at once. In 2022 I watched the same structure unwind in three days during the Luna collapse; the oracle went stale, the peg broke, the basis inverted. The architecture has matured. The reflex has not.
Rising yields help Tether's income and hurt its mark-to-market. A 33% rollover wall means the assets backing USDT reprice constantly. At the very front of the curve that's tolerable โ bills reset. But the same fiscal pressure lifting Tether's yield lifts the term premium across every duration-sensitive asset on the chain.
Bitcoin sits on the other side. It's the only large-cap asset with no coupon, no issuer, and no refinancing schedule. In a regime where the sovereign's own cost of capital reprices upward, that's not a hedge by correlation โ it's a hedge by construction. The 2024 ETF flow data I tracked for weeks showed the mechanism cleanly: institutional creation baskets react to settlement-window supply, not retail sentiment. That 15-minute lag between OTC desk sales and spot ETF purchases is the fingerprint of a hybrid market where TradFi settlement clocks and crypto volatility share one order book. That overlap is where mispricing lives.
Watch the demand, not the headline. A $39 billion auction "clearing" tells you price, not conviction. Steady demand can mean real money. It can also mean primary dealers absorbing supply they don't want and parking it. The distinction shows up later, as a tail, as a concession, as a steepening that won't reverse.
This is where I earn my keep. As an options strategist, I don't express a fiscal-dominance view by buying spot. I express it through the curve. Long-end steepeners, put skew on rate-sensitive crypto proxies, and cheap convexity on the 10-year tail. The retail book buys calls on Bitcoin because "debasement." The desk buys the hedge that pays when the auction tails. The trade isn't bullish or bearish on crypto โ it's long volatility on the fiscal calendar.
Arbitrage is just efficiency with a heartbeat. And right now the heartbeat is the auction calendar.
Everyone is trading the jobs number. Weak September payrolls, cut expectations, duration bid. Clean story.
It's the wrong variable.
The market is pricing the front end while the back end is repriced by a fiscal calendar no central bank controls. The blind spot: a rate cut does not lower the refinancing wall. If Treasury must roll a third of its debt into a market demanding 5%+, the marginal buyer sets the price, not the FOMC.
Retail sees "cuts coming" and buys duration. Smart money sees auction tail risk and positions for the opposite โ long-end steepening that survives dovish pivots. And the stablecoin crowd believes it holds cash. It holds the front end of the same trade, un-audited, at a spread it cannot see.
I'll admit a scar here. In late 2025 I handed $50,000 to an AI agent to run options strategies. It overfit to a calm volatility regime and lost 60% in three weeks when a single policy headline broke the pattern. The lesson wasn't that models fail. It was that models fail exactly when the regime shifts โ and we are in a regime shift now. Human-in-the-loop isn't a preference. It's a risk control.
Watch two levels. On the 10-year, 5.5% confirms fiscal dominance; 4.5% is the fade. Watch bid-to-cover on every long auction โ a visible tail is the tell that "steady demand" is dealers catching supply, not conviction. Position before the print, not after.
The question for next quarter isn't whether the Fed cuts. It's whether the bond market lets the cut matter.