The catalyst claim that grew in transit

Lila Sciences, a Flagship Pioneering venture with $550M raised, runs autonomous labs where one AI picks recipes of chemical elements, another reads the experimental results and proposes the next batch, and the loop runs without a human in the middle. Within a couple of months of switching one such lab on, the company says, it surfaced electrode compositions built from cheap, abundant elements that perform comparably to platinum for the oxygen-evolution reaction — the step in green-hydrogen production that today leans on iridium, one of the scarcest metals on Earth.

If real, that matters: iridium scarcity is a hard ceiling on scaling the electrolyzers that split water into clean hydrogen, and a parity catalyst with no precious metal in it would lift that ceiling. The same approach, Lila says, is now aimed at rare-earth-free magnets for motors and turbines.

But follow the headline number. The CEO, on the record in an April 2025 interview, said the catalysts were 1,000x cheaper than platinum. By the time the claim reached the Abundance Summit stage and a newsletter a year later, it had become 10,000x — a tenfold jump with no fresh experiment behind it. There is no paper, no preprint, no third-party benchmark, and no released figure for how much current the catalyst carries or how long it lasts. The peer-reviewed work in the same field over the same window tops out at cutting iridium use by about 80%; nobody has shown the full elimination at parity that Lila asserts but won't publish. The expensive, telling part of materials discovery is not having the idea — it's synthesizing the thing and proving it holds up, and that is precisely the part still unseen.