The editorial argues that JetBrains' two-decade formula of selling best-in-class language-specific IDEs via annual subscriptions just hit its first trend break, and the cause is AI-native editors like Cursor, Windsurf, Zed, and Claude Code peeling off the premium professional tier. The hierarchy is being dismantled from the top, not the bottom, which is historically where JetBrains was most defensible.
The piece emphasizes that senior developers — the cohort most willing to pay $200/year for correct refactors and reliable indexing — are the ones defecting to Cursor, Zed, and terminal-first agents like Claude Code and Aider. This inverts the usual disruption narrative where cheap/free tools erode from below; here the premium segment is being stripped first.
By surfacing the Helgi Library financial data to Hacker News and framing it as the 'first time in tracked history,' the submitter positions this not as a minor fluctuation but as a signal event for a company previously known as quietly one of Europe's most profitable dev-tool vendors. The 100-point score suggests the HN audience validated this framing as noteworthy rather than noise.
JetBrains — the Czech-headquartered maker of IntelliJ IDEA, PyCharm, WebStorm, Rider, GoLand, and the Kotlin language itself — just posted its first net financial loss in the entire window Helgi Library has tracked the company. For a privately held firm that spent the 2010s being casually described as 'quietly one of the most profitable developer-tools companies in Europe,' that is not a rounding error. It is a trend break.
The raw numbers matter less than the shape of the curve. JetBrains spent twenty-plus years compounding revenue off a dead-simple formula — annual subscriptions to the best language-specific IDEs on the market, sold one seat at a time, mostly without a sales team — and that formula just broke for the first time. Revenue growth has been decelerating across the industry's premium tooling vendors, but JetBrains was supposed to be the one that didn't flinch. Kotlin alone gave them a decade of structural tailwind inside the Android ecosystem. The Rider/.NET bet paid off. Fleet was a reasonable hedge.
And then the AI-native editors showed up.
For most of the last fifteen years, the IDE market had a clean hierarchy. VS Code owned the free tier and the long tail. JetBrains owned the paid professional tier — the people who would expense $200/year because a working debugger, a correct refactor, and an index that didn't lie were worth it. Everything else was noise.
That hierarchy is being dismantled from the top, not the bottom. Cursor hit a reported $500M ARR run-rate in its second year by forking VS Code and wrapping it in a coherent agent experience. Windsurf (now Codeium's editor) made the same bet. Zed, backed by the people who built Atom and Tree-sitter, is shipping native collaboration and a thoughtful AI panel on a Rust core. Claude Code, Aider, and the rest of the terminal-first agent crowd are peeling off the senior-dev segment that used to be JetBrains' most loyal cohort — the ones who paid without arguing.
JetBrains' response has been AI Assistant and, more recently, Junie, their coding agent. They are credible products. They are not, by most reports from working developers, the ones people are switching *to*. The core architectural problem is that JetBrains' value prop was built on the deepest static analysis in the industry — a decade of PSI trees, inspection engines, and language-server-before-LSP tooling. That moat is worth less when the thing on the other side of the screen is an LLM that will cheerfully refactor a codebase it has never indexed, correctly, most of the time. The marginal value of 'perfect symbol resolution' drops when 'good enough resolution plus a model that reads intent' is a plugin install away.
The community reaction on Hacker News and in JetBrains' own forums has been telling. The top-voted comments aren't gloating — they're variations of 'I still use IntelliJ for Java, but I write everything else in Cursor now.' That is the quiet death of a bundle. When a single developer uses three editors in a week, the per-seat all-languages pack stops making sense. And JetBrains' bundle pricing — historically their best unit-economics lever — becomes the thing that pushes churn.
There is also a cost-side story that deserves its own paragraph. Running an AI assistant competitively means paying frontier-model API bills, or training and hosting your own. Cursor and Windsurf take a markup on inference and have raised capital explicitly to burn on it. JetBrains, as a bootstrapped, profit-funded company, has historically avoided that kind of spend. Competing with VC-subsidized inference while running a profitable P&L is the kind of fight that shows up on an income statement one quarter later, and that is roughly what the numbers suggest just happened.
If you run engineering at a company that currently pays for the JetBrains All Products Pack per seat, this is the quarter to actually look at the renewal. Not to cancel reflexively — IntelliJ IDEA Ultimate is still the correct answer for a serious JVM shop, and nothing in the agent world has displaced it for debugging a Spring Boot deadlock. But the all-languages bundle for a mixed team is harder to defend when half your frontend devs have already moved to Cursor on their own time and are expensing it individually.
The practical move is to segment. Keep JetBrains seats where the IDE's language intelligence genuinely matters — JVM, .NET, large Kotlin or Scala monoliths, Rider for Unity — and move the rest to a native-AI editor plus a per-seat agent subscription. You will probably spend the same money. You will likely get more leverage out of it, because the agent is the thing writing code now and the editor is just the surface.
For individual developers, the signal is softer but real: the IDE lock-in you've been carrying since 2015 is worth less than it was. If you've been meaning to try Zed or Cursor and the muscle memory is the only thing stopping you, the market is now actively subsidizing your switch. Keybinding parity has gotten shockingly good.
None of this means JetBrains is going away. They still own Kotlin, they still have the deepest static analyzers in the business, and a first annual loss on a decades-long profit streak is a wake-up call, not a death notice — plenty of category leaders have taken one bad year and come back meaner. The interesting question is whether the response is to double down on Junie as a first-class agent inside IntelliJ, or to open the platform further and let the agent war happen on top of their runtime. Either path is defensible. The one that doesn't work is treating this as a cyclical dip. The editor layer is being repriced in real time, and the companies that acknowledge it first are the ones that will still be charging for seats in 2030.
All the death knell comments - Is no one looking at the revenue line? Revenue still trending the same. Costs presumably haven't skyrocketed. They've invested in something big. Thats probably a good thing, and often needed to evolve.
Hopefully they're investing in better llm integration. I really want to automatically use the jetbrains diff for every llm code change locally.I also want to have the jetbrains code quality scans automatically ran and fixed, and maybe something that enforces permissions and a sandbox. Basically
A bit tongue-in-cheek but I'm amazed by people that don't open an IDE or look at code. Are you paid to press enter at anything that Claude present to you?Personally I usually use Claude inside the Jetbrain's IDE terminal, even if I code by hand obviously less. I still like using the d
As a very heavy user of IDEA, DataGrip and PyCharm, I still appreciate the level of power and flexibility JetBrains offer but continue to be dismayed by the absolutely massive binaries of their IDEs and really bad performance, especially their file watching capabilities which take ages to sync the c
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
At the beginning of the AI race, JetBrains had:- top-tier language intelligence- all sorts of applications for building software, from IDEs to task trackers- a huge, loyal user base- being big enough to have deals with leading AI providersThey had the potential to lead the AI devtools race. Right no