The main reason is that languages like Python are higher-level and require less boilerplate (e.g. types). Startup success often depends on being first, so this is essential.
But note that statically typed languages like Java are not just faster (although that's part of it). They're also easier to scale. When more developers join and the codebase grows, most of the time you'll be changing code you're much less familiar with than you are with your personal projects. Static types help because:
(Java is not the best example since it's type system isn't great. There are other languages that have much more expressive type systems, so they catch many more errors at compile time).
j
stuff ;)
because Script-Languages have a faster development cycle, they are less efficient though.
So as the company grows it makes sense to scale by gradually replacing the software by compiled languages.
But I have to point out that a lot of big companies still use ruby,python or php and just replace their weakpoints. Sentry for example runs on python but for the analysis of sourcemaps they use rust.