Nuxt vs Next, revisited: seven years later.
We wrote this comparison in 2018. Nuxt and Next have both shipped two major versions, an entire rendering model shift, and a complete rethink of how data fetching works. Time to revisit.
In 2018 we wrote about choosing between Nuxt and Next. The frameworks were genuinely similar — server-rendered Vue or React, file-based routes, opinionated conventions. Seven years on, they've both transformed, and the answer to 'which one' is now more interesting.
What Next did
Next pivoted hard. The App Router landed, server components became the default, and the mental model fragmented — server components, client components, server actions, edge runtimes. Powerful, but the learning curve got steeper. Vercel's commercial momentum is enormous, but the framework is a different animal than it was.
What Nuxt did
Nuxt 3 launched in late 2022 on top of Nitro — a deployment-target-agnostic server engine that's quietly one of the best things in the JS ecosystem. Nuxt 4 stabilised the directory conventions, kept the rendering model coherent, and didn't fragment the mental model. Same Vue you wrote last year. Same Composition API. Cleaner build.
Where each one stands, mid-2025
| Concern | Nuxt 4 | Next 15 |
|---|---|---|
| Rendering model | SSR + SSG + ISR, unified | App Router + server components, fragmented |
| Server engine | Nitro (deploy-anywhere) | Vercel-native, others via adapters |
| Routing | File-based, layouts nest | App Router with parallel + intercepting routes |
| Data fetching | useFetch / useAsyncData | Server components + use() / fetch() |
| Learning curve | Gentle | Steeper after App Router |
| Best fit | Brand sites, content, SaaS, hybrid | React-heavy teams, edge-first apps |
How we now choose
- 01Default — Nuxt. The mental model survived two major versions without becoming three frameworks in a trenchcoat.
- 02Heavy React-native team and a static-export brief — Next still works, but Astro is increasingly the better answer there.
- 03Edge-heavy server logic — Nitro's preset story is calmer than Next's edge runtime quirks. Nuxt wins.
- 04Anything where the marketing team will edit content — Nuxt Content + Nuxt Studio is best-in-class. No real Next equivalent.
What we got wrong in 2018
- 01We said Next was 'closer to vanilla React'. Today, App Router isn't 'closer to' anything — it's its own thing.
- 02We said the Vue ecosystem was 'smaller but more cohesive'. The cohesion compounded. Smaller compounded the other way; Vue 3 + Nuxt + Pinia + Nuxt UI feels like one product.
- 03We undersold the docs gap. It's bigger now, not smaller.
Where 2018 felt like a coin flip, 2025 feels like a clear default — and the default is Nuxt.
We still ship Next when the brief or team demands it. But where 2018 felt like a coin flip, 2025 feels like a clear default — and the default is Nuxt.