On 28 January 2026 gold touched $5,602.22 an ounce, the highest price in its five-thousand-year history and, for the first time, a record in inflation-adjusted terms as well. It cleared the real value of the 1980 peak.
In mid-October it traded near $4,150, roughly a quarter below that peak.
Over the same stretch, nearly everything that is supposed to drive gold higher got stronger. The US–Iran war has run since 28 February and shipping through the Strait of Hormuz has been disrupted for months. Brent crude has spent much of the autumn above $100. Central banks kept buying.
Gold fell through all of it.
The usual explanation is that the market is being irrational, or that the safe-haven bid is simply exhausted. Both miss the point. The answer is in the bond market.
The variable that matters most is not the war
Gold pays no coupon. Its real competition is the return available on a risk-free bond after inflation, which is why the US 10-year real yield — the Treasury yield after stripping out expected inflation — is the single most important number for gold. FRED's series for it, DFII10, stood at 2.78% in mid-October, with the nominal 10-year above 5%.
In our model each basis point on that yield is worth about 0.11% on the gold price, so every 25 basis points is worth roughly 2.75%. Few drivers come close in size. The war, central-bank buying and the dollar all matter, but none moves the needle by that much for a quarter-point change.
And this year the real yield has been rising, not falling — with the Federal Reserve leaning towards hikes rather than cuts.
Now the uncomfortable part: oil is what moved that variable
Here is where the conventional wisdom inverts.
Expensive crude reaches the gold price down two separate paths, and they point in opposite directions.
Two channels, opposite signs
The upper channel is the one everybody knows: war and expensive energy make people want bullion.
The lower channel is the one that actually set the price this year. Crude above $100 keeps headline inflation well above the Fed's 2% target. Sticky inflation keeps the Federal Reserve tightening rather than cutting. Tightening lifts real yields. And real yields, as the arithmetic above shows, are what gold trades against.
So the oil price that was supposed to rescue gold is the mechanism that crushed it. Expensive crude did not fail to support bullion. It actively financed the thing that sank it.
Where the channels cross: $120
If oil helps gold a little and hurts it a lot, there must be a level where the second effect takes over. We went looking for it.
Holding every other assumption fixed and sweeping the Brent price across our Monte Carlo model — 20,000 paths at each level, with common random numbers so the comparison is clean — the three-month median gold price traces a hump.
Gold's three-month median against the Brent price
Below roughly $120, crude is gold's friend. Above it, crude is gold's enemy. Brent is $110 — close enough to the crossover that the next escalation could push oil through the line where it stops helping.
Read the right-hand end of that curve carefully, because it is deeply counterintuitive. A Brent price of $180 — a genuine supply catastrophe, Hormuz closed, the worst geopolitical outcome on the board — produces a lower median gold price than $80 oil does. The inflation and policy consequences of that much crude overwhelm every ounce of panic buying it generates.
Anyone holding gold as a hedge against an oil shock should sit with that for a moment.
What would actually change this
The honest answer is: the Fed, and not much else.
Prediction markets put a hike at the 9 December meeting at roughly 70%, with 28 October far less likely. As long as hikes keep landing, the channel that matters most for gold stays open, and no amount of geopolitical drama offsets it. That is the lesson of the last eight months, and it is why we would treat any war-driven gold rally as something to fade rather than chase.
The turn comes when the market stops pricing hikes and starts pricing the end of them — which is a story about inflation data, not about Iran. A ceasefire helps gold mainly by lowering oil, lowering inflation and letting the Fed stop; it also removes the war premium, which is why the net effect of peace is far smaller than people assume and can even run the wrong way.
Watch the real yield, not the headlines. Every 25 basis points it comes down is worth roughly 2.75% on the gold price in our model, and you can track it live on the simulator.
Disagree with any of this? Every assumption above is a slider on our simulator. Set the Fed path, the oil price, the real yield and central-bank demand to whatever you believe, and 2,500 Monte Carlo paths re-run in your browser to show you what your view implies.
A note on our numbers
The simulator now takes its market inputs straight from FRED, Kalshi and Binance rather than from figures typed in by hand. An earlier hand-maintained build carried a wrong real yield, and on a model where that is the dominant input, the error moved the three-month median by about $230. We would rather say so than quietly reprice.
The model also has a known weakness: at horizons beyond three months, assumed central-bank demand compounds while the real-yield channel fades, so the six-month and one-year medians drift upward by construction. The simulator flags this on those tabs. We would not rely on the one-year number, and we are not asking you to.