The Fed Doesn't Set the Rate Path Anymore. A Tanker Does.

The Fed Doesn't Set the Rate Path Anymore. A Tanker Does.

Somewhere off Oman on Monday night, an unidentified projectile hit a supertanker transiting the Strait of Hormuz. By Tuesday morning, WTI was up 5.9% to $90.82, Brent had cleared $91, the 10-year Treasury yield had ripped to 4.78% — its highest level since January 2025 — and CME FedWatch had repriced September hike odds from roughly 36% to north of 65% in the space of four trading sessions. Nobody at the Eccles Building did anything. Nobody voted. Kevin Warsh didn't give a speech between Friday and Tuesday. The rate path moved anyway, because the actual reaction function driving front-end pricing right now isn't the dot plot — it's a chokepoint 21 miles wide that a handful of Revolutionary Guard fast boats can threaten on any given evening.

This is worth sitting with, because it inverts the entire architecture the market has spent eighteen months building around Warsh's tenure. The story since Jackson Hole has been about a Fed chair who wants markets to stop looking to him for the next trade — a "quieter Fed," data-dependent, allergic to forward guidance. Fine in theory. But data-dependence only works if the data is generated by an economy behaving like an economy. What's actually happening is that the Fed's reaction function has been outsourced to a war it has no ability to end, fought over a waterway it does not control, by parties who do not care what the September FOMC statement says. Warsh can be as quiet as he wants. The market isn't pricing his silence — it's pricing the compounding probability that Hormuz stays disrupted long enough to keep core goods inflation hot through Q4, and it's doing so through a transmission mechanism that runs oil price to breakevens to term premium to hike odds, with the Fed itself reduced to a downstream recipient of its own policy inputs.

Look at what's actually cracking under that transmission. The 30-year sits at 5.28%, a level that turns Bessent's doubled Treasury buyback program — $4 billion a day, running through early November — into what it always kind of was: a circuit breaker for a fuse that keeps getting relit from outside the box. You cannot buy back your way out of a structural mismatch between $40 trillion of debt and a term premium that's rising because of tanker insurance costs in the Gulf. The buyback absorbs supply; it does nothing to the demand side of the equation, and the demand side is being set right now by a Fesharaki-style shipping risk premium, not by anyone's assessment of the neutral rate.

Semiconductors are the cleanest read on how badly this discomforts the AI capex trade specifically. Intel, up 143% year-to-date, and AMD, up 120%, are getting hit hardest precisely because they carry the most multiple compression risk when the discount rate moves — Nvidia's relative resilience (still up only 19% YTD by comparison, oddly the laggard of the group) is basically an accident of already being de-rated all year. CoreWeave-adjacent names like TeraWulf are getting torched for a second, unrelated reason: a bitcoin miner repositioned as a compute provider is now simultaneously exposed to rate-driven multiple compression on the AI leg and a listless crypto tape on the legacy leg. There is no hedge inside that stock anymore. There's just correlation.

And that correlation breakdown is the real story, more than any single asset's move. Gold, which spent August doing its debasement-trade job beautifully — up roughly 10% on the month, near its record — is now falling on the same headlines that are pushing yields up, because a rate-hike narrative beats a fiscal-debasement narrative when the two collide in the same news cycle. Bitcoin is doing the least interesting thing it can do, which is nothing: parked at $78,000, "consolidating," which is crypto-speak for a market that has no idea whether it's a risk asset getting sold with tech or a monetary hedge that should be catching a bid on sovereign stress. Three assets, three incompatible stories, one Tuesday morning. That is not a coincidence of the news cycle. That is what happens when the actual driver of policy expectations — a war zone — doesn't map cleanly onto any of the frameworks investors use to organize a portfolio. Inflation hedges don't know if they're pricing supply shock or demand destruction. Duration doesn't know if it's pricing a hike or a recession. Every model in every 60/40 shop right now is running on an assumption that the Fed is still the marginal price-setter of forward rates, and for the second time this cycle — the first being the July jobs collapse — that assumption is visibly wrong.

The dangerous part isn't the volatility. Markets absorb volatility; that's the whole point of them. The dangerous part is the asymmetry of what's setting the volatility. A soft jobs report is reversible with one strong print. A closed strait is not reversible on anyone's schedule but Tehran's and Washington's, and neither side has given the market a reason to think this resolves before the September 16 meeting. Warsh inherits a rate decision this month that will look, in retrospect, less like monetary policy and more like a referendum on how long a maritime chokepoint stays hot. That's not a position any modern central banker has wanted to be in, and it's not one the market has priced with any real conviction — it's priced it with a coin flip, three prediction markets disagreeing with each other by double digits, and every asset class simultaneously hedging a different outcome. Nobody actually knows what September 16 looks like. That includes the people voting on it.