The Federal Reserve held its policy rate near 3.9% and, in explaining why inflation refuses to settle, pointed at the AI boom and energy costs. That is a central bank naming physical inputs — electrons, fuel, the industrial plumbing behind every new data hall — as the reason money has to stay expensive. And the commodity complex is now doing two contradictory things at once: the raw materials feeding the buildout are tight enough to show up in a Fed statement, while the metal whose entire purpose is to hedge monetary failure is being quietly sold into the floor.
Start with the physical logic, because it is the part that actually holds. A data center is a claim on megawatts before it is a claim on silicon, and megawatts are a claim on copper windings, transformer steel, and whatever generation sits on the margin of the grid — usually gas. That demand does not negotiate. It arrives as a load profile and competes with existing industrial offtake for the same cargoes and the same interconnection queue. When a hiking path is being justified by input costs rather than wage pressure, you are watching monetary policy get written by a supply chain. Chair Kevin Warsh was explicit that inflation control comes first and that politics does not enter the calculation, and Fed speakers have kept further tightening on the table.
| Symbol | Close | Change | Day range | 52-week range |
|---|---|---|---|---|
| XAU/USD | $3,875.33 | -8.26 (-0.21%) | $3,872.07–$3,887.27 | $3,871.19–$5,595.46 |
Bullion heard all of that and shrugged downward. XAU/USD opened at $3,882.75, eighty-four cents below the prior close — no gap worth the name — traded a fifteen-dollar range between $3,872.07 and $3,887.27, and settled at $3,875.33, off $8.26 or 0.21%. Thin, directionless, and barely worth a paragraph on its own. What makes it worth more than that is the location: this was a fourth consecutive red close, it leaves gold 0.11% above its 52-week low of $3,871.19, and it caps a slide of 3.5% over five sessions and 6.9% over thirty. The metal is 16.8% below its 90-day high.
Gold finished within $4.14 of its 52-week low in the same week the Nasdaq posted its worst month since March 2025. The haven bid never showed up.
The mechanism is unglamorous. Cost-push inflation that invites more hikes is the single worst macro cocktail for a zero-coupon asset with no industrial demand floor. Copper has a wire to go into; gas has a turbine to burn in; gold has a vault and an opportunity cost. I got this wrong twice this month — I argued for reclaims above $4,142.67 and $4,180 on the theory that supply-side inflation would drag bullion up with it, and both expired worthless. What I misread was the sequencing: the market is pricing the policy response to commodity inflation before it prices the inflation itself. My call for a close below $3,993.59 landed, which is cold comfort — it was the same thesis, just pointed the other way, and only the second version made money.
The counterpoint deserves airtime, because it is not weak. If energy and power costs are genuinely embedded in the price index rather than passing through, there is a point at which either the hiking path breaks or inflation expectations stop behaving, and gold is the fastest-repricing asset in either scenario. Equities already registered the strain — a sharp Nasdaq drawdown and the index's worst month in well over a year, driven by what the buildout costs rather than what it might earn. Historically that kind of equity stress generates at least a reflexive bid in bullion. It did not. Either the haven function is temporarily jammed by the real-rate signal, or positioning is simply too heavy to clear.
And a concession on the session itself: fifteen dollars of range and a fifth of a percent is not a verdict. It is a market resting on a trapdoor, which is a different thing from falling through it. The levels that matter are close by in both directions. A daily close below $3,850 — a round number traders are watching below the 52-week low — inside five sessions would confirm that the real-rate repricing still has room, and I would expect the physical trade to go quiet while it happens. A reclaim of $4,000 within the next seven sessions would tell me the opposite: that the inflation story has finally overtaken the policy story, and my October thesis was early rather than wrong. What would change my mind faster than any price print is a Fed speaker softening the hiking language while still naming energy as the culprit. That combination is how bullion bottoms.
So which side is mispricing this? The industrial complex is being told, by the central bank itself, that it is tight enough to drive policy. The monetary metal is trading as though the tightening will work and nothing will break along the way. Both cannot be right for long — and I am not at all sure which one blinks first.