Platinum fell to 1,676 dollars an ounce on 29 September, the lowest since August and an eight-week low. The immediate cause was familiar by then: elevated oil prices pushed up expectations of higher-for-longer interest rates, which raised the opportunity cost of holding a metal that pays nothing, which thinned investment flows.
The longer story is more interesting and has nothing to do with rates.
The World Platinum Investment Council expects industrial demand to rise 5 per cent in 2026, driven in part by the build-out of artificial intelligence infrastructure. It expects automotive demand to fall 4 per cent. Those two forecasts together produce a projected surplus of about 265,000 ounces, which would be the market’s first surplus since 2022.
For most of the past decade platinum’s story was simple. It went into catalytic converters. Cars were the market. The metal was in deficit because mine supply was constrained and every vehicle needed some.
Now it is being pulled by two customers who want different things on different timelines. Data centres, chips and hydrogen technology are growing demand. Electrification is shrinking it, and higher fuel prices are accelerating the switch, which means the oil spike is hitting platinum twice: once through interest rates and once by making electric vehicles more attractive to buyers who no longer need a catalytic converter at all.
A business with two major customer groups moving in opposite directions is in a genuinely difficult position, and it is a much more common position than the clean growth stories suggest.
The classic example is the cafe that built its weekday trade on a single office block and its weekend trade on local families. When the office empties and the neighbourhood fills, the headline revenue can look almost unchanged while absolutely everything underneath it has moved. The hours that matter change. The products that sell change. The staffing pattern that worked becomes expensive in one half of the week and inadequate in the other. The owner looking only at the monthly total sees stability, and makes decisions that fit neither customer.
The practical requirement is to stop managing the aggregate. Platinum as a single market is becoming a less useful concept than industrial platinum and automotive platinum, which have different drivers, different growth rates and different sensitivities. A producer that plans against the blended number will be wrong about both.
The second requirement is to notice which curve is early. Automotive demand is falling now and will keep falling as the vehicle fleet turns over, which is slow and highly predictable. Industrial demand tied to AI infrastructure is rising now, is less predictable, and could compound far faster than anybody has forecast.
The structural picture still points to tightness further out. The WPIC expects deficits averaging 331,000 ounces a year from 2026 to 2030, even while forecasting a surplus this year.
One year of surplus inside five years of deficit is not a trend. It is a crossing point between two customers.
