At KotlinConf 2025, JetBrains showcased Kotlin advancements including language enhancements, the maturation of the K2 compiler, tooling like Amper, strides in KMP, and a push into AI.
Eyes are on Kotlin 2.2, which is set to bring some rather useful features to the table. Developers can start getting ready for guard conditions in when-with-subject, multi-dollar string interpolation for more elegant templating, the ability to use non-local break and continue, and context parameters.
However, the change that’s likely to make an immediate, tangible difference to daily coding life is the K2 compiler now being the default in IntelliJ IDEA 2025.1.
K2 compiler becomes default, Amper heads towards alpha
This isn’t just a minor tweak; it’s a fundamental shift. The IDE will now tap into parts of the much zippier K2 compiler for its code analysis. JetBrains wasn’t shy about showing off the improvement, stating, “When we turned on K2 mode for the repo, compilation time dropped by over 40%.”
That’s a serious saving, especially when you consider this was tested on their colossal IntelliJ monorepo – that’s nearly all JetBrains projects, with over twelve million lines of Kotlin code. You can imagine the productivity boost across the board.
But this isn’t just about going faster. The K2 compiler is really the bedrock for what JetBrains calls a “steady, high-quality rollout of new language features.” Their focus is clear: make Kotlin even nicer to use, bolster its type safety, and speed up that crucial core feedback loop for developers.
JetBrains provided a sneak peek at features designed to deliver just that, including positional destructuring, name-based destructuring, beefed-up nullability, richer errors, must-use return values, and the CheckReturnValue annotation.
The team also gave a nod to compiler plugins, acknowledging they’re vital for unlocking new ways of programming. With K2, there’s more confidence in the stability of the compiler’s innards. This paves the way for a stable compiler plugin API for the frontend, making it easier for developers to extend the compiler with custom checks and code generation more reliably.
Speaking of developer experience, the JetBrains team gave us an update on Amper, their experimental build tool for Kotlin and the JVM. It’s steadily making its way to an alpha release.
The philosophy behind Amper seems to mirror Kotlin itself: make it developer-friendly. They’re promising a straightforward configuration path, solid IDE support, clear error reporting, and the flexibility to build extensive tooling around it. The ambition for Amper is to handle whatever you throw at it – multiplatform applications, backend services with Spring or Ktor, or even simple command-line tools.
Kotlin Foundation grows stronger, JetBrains announces 2025 Grants Programme
The Kotlin Foundation, which champions the Kotlin ecosystem and its wider use, is getting a boost with two pretty influential new members. Block, Inc. is coming in as a Silver Member, and Meta is joining as a Gold Member. That’s a strong vote of confidence from big industry players.
The Foundation’s Grants Programme, which has done a great job over the past couple of years spotlighting top-notch open-source contributions, also had its moment. A hearty congratulations to the 2024 winners: Konsist, Compose Rich Editor, Multiplatform Settings, Ultron, and Orbit MVI – all doing fantastic work.
JetBrains announced the Grants Programme is reopening this year. If you’re tinkering away on something impactful in Kotlin, this is your chance to apply for a bit of backing and well-deserved recognition.
Stepping into AI: Koog arrives, Mellum goes open-source
But if there was one theme that really wove its way through many announcements, it was AI. JetBrains is making a twin bet: helping developers build AI applications with Kotlin, and giving Kotlin developers AI-powered tools to help them build.
JetBrains didn’t just dip a toe in the water; they unveiled Koog, a brand-new framework for crafting sophisticated and dependable AI agents in Kotlin. And the good news? It’s being open-sourced. Koog comes with pre-built workflows and patterns that you can use as-is or mix and match, letting developers marry the smarts of LLMs with Kotlin’s clean expressiveness.
On the flip side, AI is stepping up to assist Kotlin developers. JetBrains AI is now deeply embedded in their IDEs, offering clever features like being able to chat with the AI about your codebase or ask project-specific questions and get genuinely context-aware answers.
The code completion magic is powered by Mellum, JetBrains’ very own LLM, which has been specifically designed for code-related heavy lifting. They claim it’s fast, accurate, and ready for the rough and tumble of production on large codebases. There’s even a version finetuned just for Kotlin. Mellum is also being made open-source.
As if that wasn’t enough, they also gave us a closer look at Junie, an AI coding agent that can tackle some pretty complex tasks across all sorts of Kotlin applications – from server-side and Android to Kotlin Multiplatform.
Kotlin Multiplatform: New plugin and iOS stability
We saw several major updates for Kotlin Multiplatform (KMP), including a new plugin for IntelliJ IDEA and Android Studio, which should help smooth out the multiplatform development workflow.
Compose Multiplatform for iOS has now reached stable status. Developers now have everything they need to build polished, responsive apps on iOS – complete with native-like scrolling, iOS-native text selection, drag-and-drop, variable font support, and natural gestures – all using Kotlin.
To make development even slicker, Compose Hot Reload is here. You can tweak your UI and see changes instantly, with “no restarts, no state loss.” That means faster iteration and a much smoother ride.
For those looking at broader Apple ecosystem integration, the first experimental release of Swift Export is pencilled in for Kotlin 2.2.20, which holds the promise of easier interoperability.
JetBrains supports the web ambitions of Kotlin
Compose Multiplatform for web is steadily improving, with work ongoing on performance and bringing more web APIs into the fold, pushing it closer to having all the same features as on other platforms.
At the same time, Kotlin/Wasm is advancing, offering faster incremental builds, trimmer output binaries, and a better overall developer experience. Both Kotlin/Wasm and Compose Multiplatform for web are on track to hit beta later this year.
Server-side Kotlin: Adoption soars, Spring partnership cemented
On the server-side, where Kotlin already has a strong foothold (nearly half of all Kotlin devs use it on the backend), things are also looking rosy. To build on this success, JetBrains announced a partnership with the Spring team.
The collaboration will focus on key areas like full null safety for Kotlin and Spring apps, creating official learning materials specifically in Kotlin, and speeding up Kotlin reflection with kotlinx.reflect. This is a smart move aimed at making Kotlin a top-tier choice for server-side work.
Within JetBrains’ own backend offerings, Ktor 3 arrives with some hefty improvements. This includes 3x faster I/O performance, better configuration support, and new goodies like server-sent events and WebAssembly support. Ktor’s popularity is clearly on the up, with a 37% growth in adoption over the last year.
The Exposed library, for those working with databases, has also had a complete overhaul. A rewritten core, support for new SQL concepts, and much-improved documentation and onboarding should make working with Exposed a much better experience.
JetBrains confirms official Kotlin LSP is coming
JetBrains confirmed an official Kotlin Language Server Protocol (LSP) is in development. This is key to getting Kotlin support into a broader range of development environments.
A new Kotlin extension for Visual Studio Code is already in its nascent stages, providing basics like code completion, navigation, inspections, quick-fixes, Java interop, and project import. An alpha release for the LSP is on the cards for later this year.
KotlinConf 2025 felt like more than just a series of updates and shows a language and ecosystem that’s not just innovating but doing so with a real sense of purpose and a clear understanding of what developers need.
See also: I/O 2025: Google arms developers with fresh AI models and tools
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
