Every 6 months this question resurfaces, and I keep flip-flopping on my answer. So I sat down last weekend and actually tried to think it through instead of tweeting a hot take.
Here's where I landed. Fair warning: I've been shipping React apps for a few years now, so I'm biased. But I've also spent evenings this year with Svelte 5, Solid, and (briefly) htmx, so it's not pure fan mail.
The short answer
Yes. But not because React is technically the best.
React in 2026 is a weird beast. The library itself has become... calm? RSC-plus-Server-Actions is finally not scary to reach for. The Compiler removed 80% of the reasons I used to yell at useMemo. And "just useState and lift it up" still works fine for a lot of apps.
Meanwhile the actual DX has quietly gotten better without anyone throwing a parade about it.
What's actually good right now
- The Compiler. If you haven't turned it on in a small project yet, do it this week. It's the first time in years I've felt like the framework was on my side about performance rather than making it my personal problem.
-
RSC in Next 15 and Remix v3. Server components stopped being a debate topic and became a tool. You either need them or you don't. The mental model finally clicked for me when I stopped trying to think of them as components and started thinking of them as "the render step that used to live in
getServerSideProps." - The docs. react.dev's tutorial holds up better than most paid courses I've seen this year.
What's still annoying
- Choosing a router in 2026 is somehow worse than in 2022. Next App Router, Remix, TanStack Router, React Router v7, Wouter, expo-router... every choice has an asterisk.
- Testing story is still a coin flip between Vitest + RTL and Playwright component testing. Nobody I know is happy with either.
- "Just use a framework" is the official answer for new apps, which is fine until you're the one who has to explain to a client why their static marketing site needs a server.
The real reason I'd still say learn it
Jobs. I hate that this is the answer, but it is. Look at any board — Wellfound, LinkedIn, We Work Remotely — and React postings still outnumber every other frontend stack by a wide margin. It's the safe default at agencies, at startups, at basically every enterprise that hires JS devs.
If you learn React well — "well" being hooks, state colocation, Suspense, at least one meta-framework, and enough server work to not be scared of a Node process — you can walk into most frontend interviews and hold your own. That's not true for Svelte or Solid yet, no matter how much I want it to be.
What I'd learn alongside React
- TypeScript. Non-negotiable. React without TS in 2026 is like PHP without an editor.
- One meta-framework, deeply. Next or Remix. Pick one and learn its data loading model until you dream in it.
- A backend runtime. Node, Bun, or Deno — doesn't hugely matter, but React devs who can't set up a proper server look junior real fast.
- The platform. Fetch, streams, cookies, cache headers. The best React devs I know are secretly just really good web devs who happen to use React.
What I wouldn't sweat
- Signals. If they come, they come. Preact signals and the TC39 proposal are cool to read about but not something I'd change a job over.
- The "React is dying" discourse. It comes around every 18 months. It'll come around again.
- Whether to use Zustand or Redux or Jotai or nothing. Pick one, ship an app, revisit in six months. This is not the decision that will make or break your project.
So
If you're brand new: yes, learn React. Do the react.dev tutorial cover to cover, build one boring app you'll actually keep online for a year (a personal dashboard, a link tracker, a workout log), then pick a meta-framework and rebuild it there. That's a two-month plan that will teach you more than any paid course.
If you already know it: I don't think you have to jump to anything new to stay employable. You can keep sharpening — the compiler alone deserves a weekend — and you'll be fine.
What I'd love to hear: has anyone made a real career move away from React in 2025/26 and been happy with it? Where did you land, and does the pay hold up? Because everyone I know who "left React" quietly came back for a contract.
Top comments (1)
Determining the technology stack is a foundational decision that should precede project scoping. While React offers an intuitive learning curve, high performance, and comprehensive web integration, it should not be considered the default solution for every SaaS product. While moving away from React is not under consideration, evaluating its suitability as a primary choice for every project remains a valid subject for discussion.