There's a story that's been told about American industry for the last fifteen years that goes something like this: we invent things, China manufactures them, and over time China ends up owning them. The story is told a lot at conferences and in op-eds, but I think the version that gets told is usually softer than the actual evidence warrants. Let me put the actual evidence in one place.

Solar was invented at Bell Labs in 1954. The first practical silicon photovoltaic cell was American. The patents were American. ARCO Solar was the first manufacturer in the world to hit one megawatt of annual production, in1980, in California. In 2000, the United States made about 22% of the world's solar panels. Today the US makes about 1%. China makes more than 80%. The IRA has poured tens of billions of dollars into trying to rebuild domestic solar manufacturing and has so far moved the needle by a few percentage points.

Lithium-ion batteries were invented by Goodenough, Whittingham, and Yoshino —two Americans and a Japanese researcher, all working on what would become Nobel Prize chemistry in the 1970s and 1980s. The first car-scale manufacturing was American and Japanese. Today China makes roughly three-quarters of the world's cells, plus most of the upstream cobalt, lithium, and nickel refining. CATL alone has roughly 37% of the global EV battery market. The IRA has been trying to rebuild domestic battery manufacturing since 2022, and while there's been real progress, the upstream supply chain is still mostly Chinese and likely will be for the rest of the decade.

In hydrogen, the trajectory is on the same pattern but compressed. In 2018, Chinese firms held about 5% of global electrolyzer manufacturing capacity. By2024, the number was about 60%. That's a faster compression of the timeline than we saw in either solar or batteries, and it's happening in a category where the underlying chemistry, which is to say the membrane is still largely American.

So the pattern, as far as I can tell, is genuine. We're two for three on losing platform-layer technologies in the broader electrification stack. The third one is in active competition right now, and the question is whether anything is different this time.

Something is different. The reason is hydrogen's relationship to defense.

Solar produces electricity. Electricity is fungible. The Department of Defense can buy electricity off a Chinese-made panel and the electricity does the same job as electricity from any other panel. The same is true of batteries — the DoD has been quietly using Chinese cells for years, because there isn't really a domestic alternative at scale and the cells just work. Solar and batteries got lost partly because there was no force in the American customer base that required them to be made domestically.

Hydrogen is different in a structural way. Hydrogen is not just a way to make electricity. It's a fuel, an industrial feedstock, a chemical reductant, a propellant, and the input to synthetic aviation fuel. The military uses it for long-endurance unmanned aircraft, forward-deployed power systems, submarine atmospheres, naval auxiliary systems, and synthetic jet fuel for theaters where the delivered cost of conventional fuel is hundreds of dollars per gallon. These are missions where the chemistry has to be domestic, not for sentimental reasons, but for the same reason the DoD doesn't buy its fighter jet engines from China.

The economics of this matter. The fully burdened cost of jet fuel at a CONUSair base is around $2.40 a gallon. The fully burdened cost in a contested or landlocked theater, including the cost of fuel convoys and force protection, can be $400 a gallon. Synthetic aviation fuel made from electrolytic hydrogen at scale costs roughly $9 a gallon. From the DoD's perspective, paying a premium for sovereign synthetic fuel is good economics, not a sacrifice, because it's competing against $400 fuel in the operationally important case.

The implication, which I find clarifying, is that this is the easiest of the three platform fights to win, not the hardest. We're not trying to out-manufacture China on a commodity. We're defending a single chemical substrate in a domestic supply chain, with a customer base that's structurally American and structurally not price-sensitive on sovereignty. The fight has fewer fronts and a more defensible customer floor than either solar or batteries did.

The historical pattern of American industrial wins, when you look at it, isn't about subsidies. It's about procurement. GPS didn't get built by subsidizing GPS manufacturers; it got built because the Air Force bought it. The internet didn't get built by subsidizing networking companies; it got built because DARPA bought a network and made it usable for academics, who then made it usable for everyone else. Jet engines, semiconductors, integrated circuits, and high-end machine tools all followed the same pattern: government as anchor customer, technology improving on the back of the procurement curve, commercial spillover happening once the cost curve was compounded enough.

That's the playbook for hydrogen, and I think it's the only playbook that'sever worked for the US in a strategic industry. Subsidies are politically vulnerable; every administration changes them. Tariffs are slow and often counterproductive. Procurement is durable, because the underlying demand —defense spending — is the most stable line item in the federal budget. If you can get the substrate onto enough defense programs to drive a learning curve, the commercial economics follow.

This is the part of the story that I think is most worth being explicit about. We didn't lose solar because we were technically bad at solar. We lost it because there was no force in the customer base that required it to be made domestically, and the manufacturing economics were unforgiving. The same is mostly true of batteries. In hydrogen, there is such a force, and the membrane sits inside it. A US Navy submarine will never use a Chinese electrolyzer, or an electrolyzer with Chinese components in it. That's why we still have a shot at keeping this one, and it's why the company looks the way it does — anchored on defense, expanding into industrial, building manufacturing capacity in the middle of the country where the customer base and the supply chain still exist.

Whether the country actually keeps it is a different question, and I'm not certain about the answer. But the structural conditions are unusually favorable, given how rare it is for the third platform fight in a stack to still be winnable when the first two have already been lost. If we miss thisone too, we'll have to write a different essay about why.

‍