DEV Community

Why Rolldown-Vite Is Going To Replace Vite (And Why You Should Care)

Richard Oliver Bray on June 11, 2025

It feels like just yesterday I was writing about replacing Webpack with Vite, and here we are again with another potential game-changer in the Java...

Read full post

Best Codes • Jun 13 • Edited

rolldown-vite is not replacing vite. Vite is just migrating from Rollup to Rolldown.

This is very clearly stated in the project README:

rolldown-vite is a fork of Vite that uses Rolldown instead of Rollup & esbuild. Rolldown is a Rust-based JavaScript bundler designed for maximum performance.

It can be used as a drop-in replacement to test out Rolldown with Vite. Eventually, this package is not needed anymore as all changes will be merged into the main Vite repository.

And just above that, there is a note which says “Temporary package.”

(Chrome highlight link to the above)

david duymelinck • Jun 12

On the Vite website I found the announcement
So instead of replacing Vite, like you mention in the title, it is just the next generation of their build engine. The fact that they had two build engines was never going to last. So now they made the choice for the future.

The elephant in the room for me is, why do we need faster and faster bundlers? I thought the time of indiscriminate use of javascript in the frontend is over?
I even seen people who adopted a no build attitude. That might be too extreme for some types of websites, but for other types it could be the default.

Richard Oliver Bray • Jun 12

I agree to a certain extent, build tools feel like extra work for a simple project. But for large scale projects they're really useful. I haven't worked on a large project in years but I remember the days working on an app 100s of components and pages, and complex workflows. Running Cypress tests via CircleCI took a while and this would have been useful back then.

Alois Sečkár • Jun 12

The next generation of Vite will be much more than just a "faster bundler". They are building a whole unified and optimized stack of JS tools - check oxc.rs/

And why? Because there is always a room for improvement. Tens of secods are already great compared to tens of minutes of compiling big enterprise Java apps for example. But units or even fraction of seconds are even better.

david duymelinck • Jun 12 • Edited

The next generation of Vite will be much more than just a "faster bundler"

That is possible, but with my comment I only addressed the jobs of the libraries that are being replaced. And that it is not a new tool you need to learn, as suggested in the beginning of the post and by the title.

there is always a room for improvement

I am not denying that. My question is alluding that if you have faster bundlers the size of the code you push to the browser is going to feel less significant. It is the same trap than having a powerful computer and running the application on that machine, to find out on other machines it runs slow.
Faster development tools can create blindspots.

Nevo David • Jun 12

Pretty cool seeing the speed moves - always gets me thinking how much tooling's changed since Webpack days.

Dotallio • Jun 12

The speed numbers you shared are honestly wild, especially for those huge projects. Have you hit any issues with existing Vite plugins not playing nice with Rolldown-Vite yet?

NDREAN • Jun 13 • Edited

I tried the plugins below without problem. The compile time of 3MB went from 2.5s down to 0.7s in dev mode, and from 15s to 3s to build the prod code, which seems like immediate.

import { VitePWA } from "vite-plugin-pwa";
import solidPlugin from "vite-plugin-solid";
import wasm from "vite-plugin-wasm";
import { compression } from "vite-plugin-compression2";
import { viteStaticCopy } from "vite-plugin-static-copy";
import { compress } from "@mongodb-js/zstd";

Kinanee Samson • Jun 13

Somehow vite has managed to outdo the competition and is now outperforming themselves. Impressive. Nice take