Skip to content

Next.js 16.4 recommends Cache Components for every app

The 'use cache' model closes its remaining performance gaps, Turbopack's disk cache shrinks by a fifth, and the Rust React Compiler cuts memory use by 30%.

Source: Next.js

Next.js 16.4 shipped on 6 October, and it makes Cache Components the framework’s recommended choice for every app rather than just greenfield projects. The 'use cache' model lets you mark individual components as cacheable while the rest of the page still renders at request time; Vercel had held off recommending it universally because some apps couldn’t match the previous model’s cost and performance guarantees. This release closes that gap, and new apps created with create-next-app now get it enabled by default.

The release also brings changes that land on every app, whether or not you adopt Cache Components. Turbopack’s disk cache now uses 20-25% less space thanks to Zstandard compression, and memory-allocation work from the Turbopack team cut the experimental Rust React Compiler’s memory use by 30% and its compile time by 15%. A new next upgrade --agent command hands a coding agent version-specific migration guides, codemods and verification steps for the upgrade itself.

Why it matters: if you held off on Cache Components over the cost caveats, that’s the gap this release targets. Turning it on is two config flags:

const nextConfig: NextConfig = {
  cacheComponents: true,
  partialPrefetching: true,
};

Everyone else gets the disk cache and compiler savings without touching any config.

Share

Related reading

ML for Web Devs

Running a model in the browser, and when that is a good idea

Transformers.js and WebGPU put real models on the client. The demo is easy; knowing which features belong there is the part worth thinking about.

ML for Web Devs

Machine learning for web developers: the only mental model you need

You do not need calculus to work with models. You need one idea, that a model is a function you fit instead of write, and the rest follows from things you already know.

Engineering

The N+1 you cannot see

Everyone knows to eager load. The queries that actually take sites down are the ones hiding in an accessor, a Blade partial, or a policy. There is a one-line switch that finds all of them.

More from The Wire

All briefs

Anthropic cuts distillation attacks on Claude by up to 99%

New safeguards, including a reasoning classifier and stricter account limits, now block most attempts to copy Claude by training on its output. Blocked requests in high-confidence cases are billed, and some integrations may need updating.

via Anthropic