The State of JavaScript 2025 survey results are in, and for the first time in years, the story isn’t about revolution. It’s about consolidation. The JavaScript ecosystem has grown up. Winners have been picked. The dust is settling. And if you’re running a development team or planning a product build in 2026, these results should shape your decisions.
Here’s what actually matters from the data, and what it means for your next project.
TL;DR
- TypeScript has won decisively: 40% of developers now write exclusively in TypeScript, and native type annotations are coming to JavaScript itself
- Vite has dethroned Webpack as the build tool of choice, with 98% satisfaction vs Webpack’s 26%
- React remains dominant at 83.6% usage, but growing dissatisfaction with Next.js complexity signals a shift in the meta-framework landscape
- AI-assisted development has gone mainstream: Claude usage doubled to 44%, while Cursor more than doubled to 26%
- The ecosystem is stabilising, which means development teams can make confident, long-term technology bets
TypeScript Isn’t Optional Anymore
Let’s start with the headline: 40% of respondents write exclusively in TypeScript. Only 6% stick to plain JavaScript. That’s not a trend. That’s a settled argument.
What’s more telling is that the number one pain point reported by JavaScript developers is the lack of static typing. The people who haven’t switched to TypeScript still want what TypeScript offers. The TC39 proposal for native type annotations in JavaScript (think TypeScript-style syntax, stripped at runtime) received 5,380 votes. The community has spoken.
For development teams, the practical takeaway is clear: if you’re starting a new project in 2026 and not using TypeScript, you need a very good reason. Type safety reduces bugs, improves IDE support, makes onboarding faster, and makes refactoring safer. The tooling ecosystem has caught up completely. Node.js now supports type stripping natively. Deno and Bun have supported it for ages. There’s no performance penalty, no setup headache, and no excuse.
At REPTILEHAUS, TypeScript has been our default for every project for over two years. The survey confirms what we’ve seen in practice: teams that adopt TypeScript ship faster and with fewer production issues.
Vite Has Won the Build Tool War
Webpack isn’t dead, but it’s on life support in terms of developer sentiment. While it still edges out Vite in raw usage (87% vs 84%), the satisfaction numbers tell the real story: Vite sits at 98% satisfaction. Webpack? 26%. That’s not a gap. That’s a chasm.
Turbopack, Vercel’s Rust-based successor to Webpack, has only reached 28% usage. Meanwhile, Rolldown (a Rust-powered Rollup replacement designed to work within Vite) jumped from 1% to 10% in a single year. The direction is clear: Vite’s architecture is winning, and the future of JavaScript bundling is Rust under the hood with Vite’s developer experience on top.
If your project still runs on Webpack and you’re not planning a migration, this is the year to start. The performance gains alone are worth it, but the real benefit is developer happiness. Build times that take minutes with Webpack take seconds with Vite. Hot module replacement is near-instant. Configuration is minimal. Your team will thank you.
The Framework Landscape: Stable but Shifting
React at 83.6% usage isn’t going anywhere. But the conversation has shifted from “should we use React?” to “what should we put on top of it?”
Next.js, used by 59% of respondents, generated the most comments of any project in the survey, and not all of them were kind. Developers praised its capabilities but expressed genuine concern about growing complexity and Vercel’s commercial strategy. One respondent put it bluntly: “the Next complexity has gotten absurd.”
This is creating space for alternatives. Astro continues to gain ground as a content-first framework. Solid.js maintained the highest satisfaction rating for the fifth consecutive year. Remix (now merged with React Router) offers a different mental model. The meta-framework space is the most interesting battleground in frontend development right now.
For agencies and development teams choosing a stack, the question isn’t whether to use React. It’s which layer to put on top. For content-heavy sites, Astro is increasingly the right answer. For complex applications, Next.js still delivers, but you need to be intentional about managing its complexity. And for teams that value performance and simplicity above all else, Solid.js deserves serious consideration.
AI Is Reshaping How We Write Code
The most dramatic shift in this year’s survey isn’t about frameworks or languages. It’s about how developers actually write code.
Claude usage doubled from 22% to 44%. Cursor more than doubled from 11% to 26%. ChatGPT usage actually declined from 68% to 60%, suggesting developers are moving toward more specialised AI coding tools rather than general-purpose chatbots.
This tracks with what we’re seeing across our client projects. AI-assisted development isn’t replacing developers. It’s amplifying them. The teams getting the most value from AI tools are using them for boilerplate generation, test writing, code review, and documentation. The ones struggling are trying to use AI as a replacement for understanding their own codebase.
The practical lesson: invest in AI tooling for your development workflow, but invest equally in your team’s ability to evaluate and refine AI-generated code. The productivity gains are real, but only if your developers understand what the AI is producing.
At REPTILEHAUS, we’ve integrated AI agents into our development pipeline for everything from code generation to automated testing. The results have been significant: faster iteration cycles, better test coverage, and more time for the creative problem-solving that actually requires human judgement. If you’re curious about what that looks like in practice, we’d love to chat.
Node.js Holds, Bun Grows, Deno Stalls
On the runtime front, Node.js at 90% usage is as dominant as ever. But the challenger race has a clear leader: Bun, at 21% usage and growing by 4 percentage points year-on-year. Deno sits at just 11%.
Bun’s appeal is straightforward: it’s fast, it’s compatible with the npm ecosystem, and it bundles a package manager, test runner, and bundler into a single binary. For teams that value simplicity and speed in their tooling, Bun is increasingly a viable choice for new projects.
That said, Node.js isn’t going anywhere soon. Its ecosystem is massive, its stability is proven, and recent improvements (native TypeScript support, better performance) have addressed many of the pain points that drove developers to alternatives in the first place.
What This Means for Your Next Build
The State of JavaScript 2025 paints a picture of an ecosystem that has matured. The days of a new framework launching every week and reshuffling the entire landscape are behind us. That’s good news for development teams and businesses alike.
Here’s the practical stack we’d recommend for most new projects in 2026:
- Language: TypeScript (non-negotiable)
- Build tool: Vite
- Frontend framework: React for applications, Astro for content sites
- Runtime: Node.js for stability, Bun for greenfield projects that value speed
- AI tooling: Integrated into your workflow, not bolted on
The ecosystem has spoken. The winners are clear. Now it’s about execution.
If you’re planning a build and want to make sure your technology choices are future-proof, get in touch with our team. We’ve been building with these tools long before the survey confirmed they were the right call.
📷 Photo by Patrick Martin on Unsplash
