There's a fact about iridium that I think makes its way into too few investor decks. The world produces about nine tonnes of it per year, almost all of it from South Africa and Russia as a byproduct of platinum mining. That's it. Nine tonnes, globally, annually. By comparison, the world produces about eight thousand tonnes of gold a year. Iridium is roughly nine hundred times scarcer than gold by production volume.

PEM electrolyzers eat iridium. The anode catalyst layer in every PEM electrolyzer cell contains iridium oxide, typically loaded at about 2 mg persquare centimeter of active area. A one-megawatt electrolyzer stack containsroughly half a square meter of active area per kilowatt, which means about 10 grams of iridium per kilowatt, or 10 kilograms per megawatt. At conventional loadings, the world's iridium supply could support roughly four hundred thousand kilowatt-scale stacks per year.

That number is not enough. To put it in scale: hitting realistic decarbonization scenarios for industrial hydrogen — which means retrofitting refineries, ammonia plants, and steel mills — requires hundreds of gigawatts of new electrolyzer capacity over the next two decades. The iridium supply, at conventional loadings, can't support that level of buildout even if you devoted every gram of global iridium production to electrolyzers and stopped using it for anything else, including the spark plugs and the catalysts and the electronics where it currently goes.

This is the kind of constraint that doesn't show up on a TAM slide because it isn't a market problem. It's a physical problem. There's no second supplier waiting in the wings. South Africa and Russia are the suppliers, the byproduct economics are what they are, and even if you dramatically scaled platinum mining you'd only marginally increase iridium output.

The way out of this isn't to find more iridium. It's to use less of it per stack. And the route to using less of it goes through the membrane, which I find satisfying as an example of how platform-level work creates downstream value in ways that aren't immediately obvious.

The reason conventional electrolyzers use 2 mg/cm² of iridium isn't that they need that much. It's that they need a margin against the failure modes that the membrane creates. Less iridium would work fine if the membrane underneath it blocked the hydrogen crossover that dissolves the iridium. The high loading is partially a safety factor against the substrate.

A better-designed membrane reduces the safety factor. Not directly, because the membrane doesn't catalyze anything. Indirectly, because the catalyst can sit thinner and more sparsely without losing performance, because the supporting layers are more stable, because the system as a whole is more forgiving of imperfections in the catalyst layer.

We've gotten iridium loading down to roughly 0.4 mg/cm². That's an 80% reduction. At that loading coupled with higher current densities, the world's iridium supply can support somewhere around four million stacks per year instead of four hundred thousand. That's not the difference between two adequate numbers. It's the difference between a critical material constraint that prevents the industry from scaling and acritical material constraint that doesn't.

The same logic applies to other scarce materials in the stack. Platinum loading in fuel cells is constrained partly by membrane behavior. The choice between Li-ion and Li-metal in batteries is constrained partly by separator behavior. Vanadium in flow batteries is constrained by membrane crossover, which determines how much electrolyte you need per kilowatt-hour of storage. In everyone of these cases, the critical material constraint partially traces back to the membrane.

This is what I mean when I say the membrane is a platform. The same chemistrychange reduces iridium in electrolyzers, increases stack life in fuel cells, decreases vanadium in flow batteries, and improves selectivity in CO₂ reduction. None of these are obvious from looking at the membrane in isolation. They emerge from what the membrane lets the rest of the system do.

The iridium case is the cleanest example because the constraint is the most quantifiable. There are nine tonnes a year. You can do the arithmetic. At conventional loading you can't build a hydrogen economy. At sub-milligram loadings, you can. The polymer underneath the catalyst is what closes the gap.

I think this is the kind of fact that should change how the industry talks about its own constraints. The hydrogen economy isn't going to be capped by electricity prices or by capital costs or by demand. Those are factors. It's going to be capped, in the medium term, by how much iridium you need per stack. And how much iridium you need per stack is, less obviously than it should be, a membrane question.

‍