Duranium wants to make metal in America - and turn the emissions into a product.

Here is a fact that sounds made up but isn't: the United States produces almost no primary magnesium. The metal shows up in everything from car parts to aerospace alloys, and the country that dominates it - by a wide margin - is China. If you are the Department of Defense, this is the kind of sentence that ruins your afternoon.

Duranium, a Y Combinator Summer 2025 company, was built to make that sentence less true. Its plan is not to lobby for a subsidy, or to invent a new element, or to do anything that requires you to believe in a breakthrough that hasn't happened yet. Its plan is to take carbochlorination - a metallurgical process that is roughly a hundred years old and extremely well understood - and stop wasting the parts everyone else throws away.

Carbochlorination works by combining a metal oxide with chlorine and carbon at high temperature, which converts the oxide into a metal chloride you can then reduce into pure metal. The conventional version of this vents carbon monoxide and chlorine, which is bad for both the atmosphere and the accountant. Duranium's reactor recirculates those streams back into feedstock instead. The company says this does two useful things at once: it eliminates direct emissions, and it lowers cost.

The genuinely clever part - the part that makes an investor sit up - is the byproduct. The process spins off bleaching agents, hypochlorite, which is a real chemical that real customers pay real money for. So Duranium is not asking you to price in the value of avoided carbon or a hypothetical policy tailwind. It is asking you to notice that when you close the loop, you get a second product for free, and that second product helps the first product compete on price with a subsidized incumbent halfway around the world.

Whether the unit economics survive contact with an actual factory is, of course, the entire question. But the shape of the argument is unusually clean for climate hardware: the emissions become an input, the waste becomes a product, and the whole thing is supposed to make money because it is cheaper, not because it is virtuous.

There are two ways to argue for domestic metal production. The patriotic version says the country should make its own titanium because depending on a strategic rival for it is dangerous. It's a good argument. It has also, historically, not been enough to overcome the fact that the imported version is cheaper.

Duranium is making the second argument, the one that tends to actually build factories: that it can produce the metal at a lower cost than the incumbent, with the security benefit thrown in as a bonus rather than as the whole pitch. The company projects operating margins above 50% at the world's lowest market prices, and a path to roughly 12% of the U.S. magnesium market by 2029.

Those are projections, and projections in heavy industry have a long and colorful history of not surviving the pilot plant. Duranium has proven its process at benchtop scale and is targeting a working pilot facility by June 2026 - the milestone where optimistic chemistry either becomes a company or becomes a lesson.