Ruby on Rails in 2025 - is it still worth it?

tl;dr - Yes

Long answer below.

Most of my development in the past 3 years was primarily done in Python and FastAPI for various reasons - the rest of the team knew Python better, there were API clients or binding only for that (especially true for AI/LLM-related stuff back in 2023), it was just easier or just enough. I firmly believe that you should choose the technology based on your needs and keep your options to what you already know as long as possible.

I do, however, keep some of my projects alive for quite a time now, and most of them are built on top of rails - some of them were started in Rails 6, and some are even older. Way older (I'm old enough to remember when there was no bundler and the times before the merge of Rails and merb)
And what I find comforting here is that I still have most of the original code working - there were some hiccups on the way with upgrades, but most of them were frontend related, not the logic or framework internals per se.

There are a lot of great companies that built their products on Rails, including Basecamp, Shopify and GitHub. It is highly productive framework, that in version 8.0.1 (which is the recent one at the time of writing this post) is capable of fully running with just sqlite as a database, caching backend and queue backend, there's also built-in authentication mechanism after all these years of using devise.

So, what do we get from Ruby on Rails 2025? A mature framework that has few requirements lets you be highly productive and puts the focus on convention over configuration.
Great for internal tools and MVPs, PoCs, startups, scale-ups, and enterprise solutions. It's kind of one ring to rule them all. And as for performance? Well, if you hit ruby and rails limits, I congratulate you - what you've built is probably worth so much now that you can spend some money migrating to other technologies, like Go and Rust.