The Scala Survey 2026 concluded a couple of weeks ago. You might have gotten a glimpse of the results at our Scalar booth, but if not, nothing's lost, you're at the right place!
If you'd like to take a look at the raw numbers, these are available at scalasurvey2026.virtuslab.com. For reference, you might also want to take a look at the 2023 results. Alternatively, we've prepared a short summary below.
Let's take a look at the state of the Scala ecosystem now and compare it to what has changed over the past 3 years.
How many Scala developers are out there?
First off, how many answers did we get? This year, the survey was completed by 1056 people, compared to 1387 in 2023. Unfortunately, that's a 24% drop. This might be due to changing social media habits, and hence the audience we can reach. But for all we know, it's a fact that we have to accept with humility: Scala isn't as popular or fashionable as it once was. Definitely a sign for community leaders, library authors, and language designers alike.
But! We have every reason to believe that the next Scala survey will show better numbers, with new people joining our ranks. There's a lot of initiatives to be excited about in the Scala ecosystem: Safe Scala for Agentic AI; direct-style stack using Ox; Scala on WASM; cats-effect on Scala Native; ZIO blocks; and new explorations of effect systems in Kyo and yaes; Scala on the GPU with Cyfra; and that's just to name a few (and I'm sure I missed quite a lot!)
Who are Scala developers?
The demographics of our respondents reveal that most are Scala developers who use the language for both work and hobby projects. Scala developers are highly experienced: they self-describe as advanced or expert-level, with over 75% having more than 5 years of Scala experience. Only 6% are relative newcomers, down from 13% in the 2023 survey. This underlines the importance of educational initiatives, such as organizing Scala Days and Scalar, driving the dynamic Scala meetup in London, or teaching Scala as a university course.
Scala developers seem to predominantly reside in the US, Poland, Germany, and the UK (a total of 40%!). A lot of Scala developers seem to have moved from the US to Poland since 2023—welcome! ;).
Versions
The topic of Scala 2 -> Scala 3 migration has come up at least a couple of times for every Scala developer. But it seems Scala has crossed an important milestone! In 2023, the predominant Scala version was Scala 2.13, while in 2026, most of us are using some version of Scala 3:

At the same time, most of the projects we are working on have mixed codebases. When we take a look at people who use only a single Scala version, Scala 3 is also at the top, but the absolute numbers are lower:

49% of our respondents are also planning to migrate from Scala 2 to Scala 3. And since AI is quite good at both Scala versions (you just need to use the right skills), this process can be further accelerated. AI-assisted migrations are one of the use cases where AI advantage is most clearly visible.
Scala version migrations, and establishing AI-driven SDLC pipelines, are both VirtusLab's areas of expertise. The team behind Scala might be just the answer to your needs.
Perhaps unsurprisingly, almost everybody is writing Scala on the JVM. But, both Scala.js and Scala Native see improved adoption, with a quarter of us using the JS-based variant, and 15% are experimenting with Scala Native:

Frameworks & libraries
And now for the main course: what every developer likes to talk extensively about, that is, their favorite languages and frameworks.
The com.lihaoyi ecosystem sees steady adoption and interest: about 40% of our respondents either use or plan to use the libraries. There are also clear educational opportunities, with another 40% unaware of com.lihaoyi's functionalities.
Typelevel keeps the reign as the most popular set of Scala libraries. Cats-core sees a steady 80% of users or prospective users. Almost everybody in Scala has heard about cats, cats-effect, or http4s.

ZIO, a competing functional effect system, sees lower awareness of its offering, with about 50-60% of survey participants being either active or future users. This is also on par with 2023 numbers. Compared to cats, a larger proportion of users are not interested in that approach to developing Scala. 30% are not interested in ZIO, while about 15% would not want to use cats.
Our own VirtusLab ecosystem sees mixed adoption. While a large portion of developers use or would like to use sttp-client & Tapir (70%—thank you! :) ), we've got some work ahead of us to extend the offering, develop integrations, and cover more use cases with direct-style Scala. Only about 50% have made up their minds about Ox. The other 50%—we're coming with an Ox-themed talk to your local meetup or conference!
Among the frameworks we have surveyed, Akka/Pekko shows the strongest negative sentiment, with over 40% of respondents not being interested in the offering. At the same time, almost everyone has heard of Akka/Pekko, and the other 40% are happy users delivering production value every day!
JSON, IDE, tooling
The absolute highlight of any Scala discussion is JSON libraries. That's why this year, we have a dedicated question covering only that area.
Who's the winner? Circe! It sees an adoption of about 70%, and is leagues ahead of the competition.

When it comes to IDEs, IntelliJ is also a clear winner—80% of us use it for Scala development work, with Metals-based editors being a close second (70% use some Metals-based editors). Which also means, we very often use more than one IDE for our work! By the way, we have only slightly changed our habits since 2023.
Finally—build tools. sbt remains the de facto standard Scala build tool. Everybody has heard about it, 80% are using it, 15% are staying away. But, scala-cli is seeing stronger adoption, with about half of the participants using it, and another 30% eager to learn. Mill keeps a strong interest (many people would like to try), but only 20% are actively using it. Bazel has even fewer users, but a significant number who would like to learn more.
But tooling does not end here. We all use and love scalafmt—and right so. It's a great example of a focused tool that does one job well. It also pre-empts any code formatting discussions. It's high time we put that behind us, and use a single "good enough" style—there are other, more interesting areas to talk about.
What's next?
We've also asked for direction: which areas do Scala developers consider most in need of attention? What would they like to see improved?
The results aren't directly comparable to 2023, as we've broken down some categories to be more fine-grained. That said, tooling in general remains the #1 concern, closely followed by compile / build speed.
As a side comment, I have the feeling that the compile speed is already quite good, but where improvements would be most welcome is build speed. Some of the build overhead might actually be gone with the introduction of sbt 2. But looking at the almost-instant incremental compilation when working with Metals, there's further room for improvement in that area. And this is mirrored in the third most popular suggestion answer: build tools.
Check out the work on speeding up CI in large codebases. Krzysztof Romanowski's & team are experts in this area.
Maybe equally, or even more, interesting is an analysis of the free-form "additional comments" question, where people can leave their own suggestions and feedback. It's worth keeping in mind that each category mentioned here is based on 30-40 responses, so about 3% of respondents. However, the question was totally voluntary, so it's expected that only a small fraction of people filled this in.
The top concern is consistent with the multiple-choice "what to improve" question: tooling. Sentiment analysis shows urgency and some frustration with the subject, especially around stability and performance of the IDE plugins. An interesting suggestion is for the Scala team to officially recommend a build setup—taking inspiration from Rust's cargo.
Another group of concerns relates to the job market, Scala's adoption (which we've covered in the introduction). That's, of course, understandable, compounded by AI-driven anxiety; suggestions included creating CTO-targeted "Why Scala" materials that focus on pragmatic business use cases and defensible "killer" use cases.
A number of people are also fatigued with the fragmentation of the ecosystem and "library overload". It is true that Scala seems to be especially well-suited to invite experimentation, probably because of its flexibility in expressing abstractions. Thanks to that, a number of concepts have been battle-tested in Scala and have now been widely adopted in the industry.
But in the end, one does need to provide an AI agent with a stable tech stack that covers common problems. As for our part, that is happening: we're consolidating around the direct-style Scala stack. Cats-effect is stable & mature. We're also maintaining the Scala Toolkit. That said, your Scala ecosystem, whether functional or direct-style, could certainly use your help: either through writing integrations, creating ecosystem guides, or publishing what works for you.
See you next time!
Just as a reminder, you can browse the full results on the dedicated website. We'd like to thank each Scala Survey 2026 participant for their time. I think it's valuable for every Scala programmer to see the bigger picture, understand what others are using, and identify potential pain points.
The age of AI is upon us, and this might be just the occasion for Scala to thrive. Short feedback loops are crucial for coding agents, and Scala, with its advanced type system, provides just that. Combined with concise syntax, you get a productive, token-efficient language.
I'm already curious what the next Scala survey will bring—who knows, maybe even sooner than in 3 years?

