Languages less than 10 years old are still in their infancy. It's fairly common for the first 10 years of a language to be quite niche.
The Languages bought by large corporations can avoid that 10 years by making it used by decree, that can make it more popular sooner, but it doesn't guarantee long term success,
There is probably a language that exists today that will be a top 50 language in a decade or two that neither of us knows about. Time is one of the best design resources.
10 years is enough time. The industry was absolutely using Perl (1987) in 1997 and I was using Python (1991) professionally by 2003.
I mentioned Zig already; Julia also feels like it fizzled in a similar way, by getting attention without commensurate adoption.
But I will concede that you're right about Lean. It feels important that it's solving tomorrow's problem (i.e. today's) rather than yesterday's problem.
Things do go a little slower these days however. It was 5 years and one month between the release of the Vic-20 and the Amiga. How different is your computer to the one you had 5 years ago?
There is a lot of news right now about AI writing proofs in Lean (2013.) Lots of people rely on Elixir (2012) and Gleam (2016) is up and coming. I wouldn't give up on Scala (2004) just yet, it's really one of the best languages out there with some great libraries and runs on the JVM. You could argue that Reason (2016) and ReScript (2020) are just dialects of OCaml but they are industrial strength tools that people rely on in production.
...and don't even try to take Clojure (2007) away from any hands that are not truly cold and dead.