Every few months somebody at 4AM writes a doc titled 'Do we still need Next?'. It's a ritual — one engineer gets cranky with a bundler error, proposes Astro, we all discuss for a week, and we stay with Next. This is the version of that doc I'd like to publish.
The shortlist
Four frameworks we seriously reconsider every year: Next.js (App Router), Astro with Islands, Remix (now React Router), and SvelteKit. Each has a genuine case for the kind of sites we build.
- Marketing-heavy pages with occasional interactivity
- Dashboards behind auth with real-time updates
- Ecommerce catalogues with 10k+ SKUs
- Blog and documentation content
Why Next, specifically
Three reasons, in order of how often they settle the argument. First: server components let us write data-heavy pages without a state-management detour. Second: the ecosystem is so large that any niche requirement — PDF rendering, image optimisation, auth — has three battle-tested options. Third: Hostinger VPS deploys a Next app in roughly the same way it deploys anything else.
“The framework you reach for when you're tired on a Friday matters more than the one that wins your benchmark.”
When we don't use it
For pure-content pages with no dynamic anything — we still reach for plain HTML and a sprinkle of CSS. For mobile-heavy product flows, Flutter wins. For internal tools with five users, plain React + Vite is usually fine.
// A minimal page with a server data fetch.
export default async function Page() {
const items = await getItems();
return <List items={items} />;
}The point of a framework isn't to win benchmarks. It's to let four people keep a platform alive without hiring a tenth. Four years in, Next still clears that bar.