Tearing apart the Russian palladium tariff trap

Palladium.

The financial press is treating this as a trade war footnote. Russian palladium, tariffs incoming, geopolitical noise. CNBC filed it next to the steel and aluminum sections. If you've spent any time inside the actual supply chain for this metal, you know that framing is wrong in almost every direction that matters.

Palladium is not a geopolitical abstraction. It is the thing sitting inside the catalytic converter of every gasoline-powered car built since about 2001. Over 80% of annual demand comes from one use case: scrubbing exhaust. There is no substitute that works at the same loadings, in the same temperature window, at the same cost — unless you are an automaker with eighteen months of engineering runway and a procurement team willing to bet on platinum price stability. Most are not.

Nornickel produces roughly 40% of global palladium supply. That is not a rounding error — that is the single largest swing factor in the market, mined out of a complex in Siberia that has been running since the Soviet era and is not going to be replicated anywhere else on a relevant timescale. The US Department of Commerce just set the anti-dumping duty on that material at 132.83%. The countervailing duty determination — which could add another 109% on top — is still outstanding. The ITC injury ruling, which triggers the whole mechanism retroactively to July 2025, is expected any day.

Here is the part the trade war framing misses entirely: the only domestic US producer, Sibanye-Stillwater's Montana operation, is structurally unprofitable at current palladium prices. The company said so last year in a public filing. Nornickel said so too. The tariff is not protecting a healthy domestic industry. It is protecting a mine that needs the price to go considerably higher before it can fund its own operation. That is a completely different kind of trap.

Numbers. No narrative.

The global palladium market runs on roughly 7 million ounces of annual mined supply. Nornickel accounts for around 2.8 million of those, with guidance already flagging a production decline to the lowest output in two decades. South Africa supplies most of the remainder — from the Bushveld complex, where PGMs come out co-mingled with platinum and rhodium and nobody is ramping for palladium specifically. There is no reserve army of supply waiting offstage.

US imports from Russia ran at 27.6 tonnes in 2024 — roughly 40% of US palladium imports by volume. At 132.83% anti-dumping duties, that material either gets rerouted to Europe and Asia (suppressing prices there, tightening the US) or stops flowing entirely. Either way, US autocatalyst manufacturers — Ford, GM, the transplant plants in Kentucky and Tennessee — now face a two-track palladium market where the domestic price decouples from the global price. That spread is a cost. It lands in gross margins on every internal combustion vehicle built in America.

The flow of consequences runs like this:

The retroactivity clause is what makes this genuinely dangerous. If the ITC finds injury, the countervailing duties trigger retroactively to July 22, 2025. Any US importer who kept buying Russian palladium during the investigation window — expecting a dismissal or a lower duty — is now sitting on a 10-month liability bill they did not budget for. Some of them are automakers. Some of them are refining intermediaries. The accounting shock from that retroactive charge lands in Q2 earnings season and nobody has modeled it.

The policy logic here has a structural flaw that nobody in the trade press is writing about. The tariff was filed by Sibanye-Stillwater to protect its Montana operation. The Montana operation is unprofitable at current palladium prices. Nornickel said publicly that North American PGM mining cannot break even at current spot levels. The company that filed the petition essentially confirmed this in its own earnings materials — it booked a 3.8 billion rand impairment on its US business last year. The tariff creates a protected market. It does not create a viable mine at the prices currently prevailing.

So the US has now restricted the cheapest available supply source while the domestic alternative remains uneconomic. The gap between those two facts is where the price has to travel to make the machine work again. That is not a policy outcome — it is a mechanism. You can either get Russian palladium at market prices, or you can get Montana palladium at whatever price makes Montana palladium viable. Those are not the same number.

The substitution valve is real but it runs on engineering time, not financial time. Platinum crossed to a premium over palladium in late 2025 — a reversal of the prior decade's relationship — which is exactly the signal that should trigger automakers to start swapping. WPIC estimates reverse substitution back toward palladium could hit 250,000 ounces by 2029. That is a forecast for three years from now. It is not a solution for Q3 2026. I have seen analysts present the substitution option as if automakers can reroute a catalytic converter formulation the way a logistics manager reroutes a shipping container. They cannot. BASF has been engineering flexible-loading autocatalysts for years precisely because the industry needs years to change loadings.

Meanwhile the recycling backstop is weakening at exactly the wrong moment. US scrap palladium recovery had been rising on elevated PGM prices through 2025, but the Chinese scrappage subsidy program — which turbocharged vehicle scrapping and thus autocatalyst recycling — is being wound down. The secondary supply cushion that helped paper over the deficit for two years is getting thinner. The market is pricing palladium in a range of $950 to $1,500 per ounce for 2026. That range assumes the supply mechanics stay roughly stable. They are not staying stable.

The market is pricing this as a US-Russia bilateral trade irritant. It is actually a forced repricing event with a hard floor set by the cost structure of a single mine in Stillwater County, Montana. The mine is not large. The floor is not low. And the ITC ruling that triggers the whole mechanism retroactively could drop any day this week.

The ITC ruling is the trigger. When it lands — and it is expected within weeks — several things happen simultaneously that the market has not priced into a single coherent view.

The retroactivity shock is the first wave. Importers who accumulated Russian palladium inventory during the investigation window — betting on a dismissal or a tolerable duty — now face an unanticipated cost that goes straight into working capital. Some of those positions are held by trading intermediaries. Some are on autocatalyst manufacturer balance sheets. None of it was flagged in Q1 earnings guidance because the determination wasn't final. It is final now.

The second wave is the supply reroute. Nornickel does not stop producing when the US market closes — it redirects. Russian palladium floods into the European and Asian spot markets, suppressing prices there while the US market tightens against a constrained non-Russian supply pool. The global palladium price becomes a blended fiction. The relevant number for US buyers is whatever South African material plus secondary recycling can clear at, which is higher. This bifurcation is the trade, and it has not been priced into any public equity that I can see.

Where the capital goes is narrow and specific. Not the broad precious metals ETFs — those are blended with gold exposure, platinum names still grinding through their own substitution dynamics, and rhodium positions that are dead weight at current prices. The edge is in Sibanye-Stillwater equity, which has been trading as a distressed restructuring story for two years and has not yet been repriced as a protected critical mineral franchise in a closed US market. Those are two very different valuations.

The risk is obvious: the ITC could rule against injury, the duties collapse, and the thesis evaporates. I'd assign that low probability given that the Commerce Department has already issued final determinations at 132.83% and 109.1% — the ITC rarely breaks the other way after those findings. The second risk is that automakers accelerate platinum substitution faster than anyone expects, destroying the demand base entirely. That is an 18-month engineering story, not a this-quarter story. The physical palladium market does not care about 2029 substitution forecasts. It cares about who has the metal in the warehouse in Columbus, Montana. Right now, the answer is increasingly — only one company.