The interesting thing about a platform technology isn't the platform itself. It's what becomes possible everywhere downstream when the platform improves.

When transistors got better, the obvious downstream effect was that processors got faster. The less obvious effects were that solar got cheaper, image sensors got smaller, MEMS devices became economically viable, and a few dozen other industries — none of which had been waiting for silicon to improve in any deliberate sense — found that they could now do things they couldn't do before. The platform improved, and the downstream industries reorganized around the new possibilities.

The membrane has the same kind of downstream surface, and I think most people would be surprised by how long the list is.

In hydrogen production, a better membrane means cheaper electrolyzers and 80% less iridium per stack, which is more important than it sounds because iridium is one of the rarest metals on Earth and the world produces about nine tonnes of it a year. Reducing the loading is what turns hydrogen from a niche industry into something you can actually scale.

In fuel cells, a better membrane means five times the durability. This is the difference between a scienceproject and a product. The DOE has been chasing a 30,000-hour ultimate durability target for years to change the economics of trucking, of stationary backup power, and of distributed generation in general.

In flow batteries, a better membrane means lower crossover, higher efficiency, and longer life, which is the difference between long-duration storage as a thesis and long-duration storage as an investable product. Current flow battery economics are constrained by membrane resistance. Reduce that resistance and you've changed the levelized cost in a way that no amount of pack-level engineering can.

In direct lithium extraction, a better membrane means you can pull lithium out of brine without solvents. This matters because the US imports almost all of its battery-grade lithium, and the alternatives to membrane-based extraction are slow, dirty, or both.

In rare earth separation, a better membrane means you can separate neodymium and dysprosium from mixed streams without the multi-stage solvent extraction process that currently makes domestic processing uneconomic. China processes about 90% of the world's rare earths, partly because they're willing to operate the chemistry at scale and partly because nobody else can do it cheaply enough. Membrane-based separation changes that math.

In black mass recycling — the recovery of cobalt, nickel, and lithium from shredded EV batteries — a better membrane lets you do selective metal recovery that would otherwise require pyrometallurgy. The waste stream from EV adoption is going to be enormous, and right now there isn't a domestic processing industry to handle it.

In CO₂ reduction, a better membrane lets you drive carbon dioxide into ethylene, methanol, formate,and other commodity chemicals using electricity instead of hydrocarbons. This isn't speculative; it works at the bench scale. What's stopping it from being industrial is efficiency, which a better membrane can enable by offering lower chemical crossover and higher product retention.

None of these markets share customers. A flow battery customer doesn't talk to a lithium extraction customer. A rare earth separator doesn't read the same journals as a CO₂ reduction researcher. None of them think of themselves as being part of the same industry. They have very different supply chains, very different regulatory regimes, and very different cost structures. The only thing they have in common is that they all require a thin polymer film that does approximately the same job: let the right ions through with the least amount of resistance possible, and block everything else.

This is what makes the membrane a platform rather than apart. If you improve a part, you improve one product. If you improve a platform, you simultaneously improve every product that uses it, including the products that haven't been built yet.

I think the reason this isn't obvious to most people is that the membrane has been treated as a commodity for so long that nobody really expects it to improve. When something hasn't improved in thirty years, it's reasonable to assume it isn't going to. So researchers in CO₂ reduction work around the membrane limitations they have, researchers in flow batteries pick their chemistry to suit the membranes that exist, and the field as a whole optimizes around the substrate rather than trying to improve it. A DOE presentation I attended in 2024 spent twenty slides talking about new catalysts and systems engineering and ended with one line on the last slide saying a better membrane would circumvent all these other problems if only it existed.

That's how stagnation in a platform looks from the outside. It looks like everybody else's progress slowing down without anyone being able to say exactly why.

The way you'd know it had ended is that several unrelated industries would simultaneously start moving faster, and the common explanation would be a substrate that quietly got better while nobody was watching. I think that's roughly where we are now.

‍