Vite is now part of Cloudflare

TL;DR

• 1

  Cloudflare bought more than Vite: Theo argues the real prize is Void Zero's ability to turn app code into infrastructure decisions, which fits Cloudflare's biggest weakness: developer experience.

• 2

  AI made deployment the new bottleneck: what used to be 30 to 100 hours of coding plus 2 to 3 hours of setup has flipped, with agents writing the app in minutes while deployment still burns hours across Vercel, Convex, Clerk, and cloud config.

• 3

  Void's pitch matches the moment: install a Vite plugin, run void deploy, and let code define DB, KV, storage, queues, and AI resources, instead of stitching together dashboards, env vars, Terraform, and scripts.

• 4

  Cloudflare's infra is strong, its DX has lagged badly: Theo says Cloudflare apps are often more Cloudflare-specific than Next.js apps are Vercel-specific, citing Wrangler, YAML, conflicting docs, and his painful T3 Chat port as examples.

• 5

  The real platform battle is shifting left and right: Vercel owns more of the framework-to-compute stack, Cloudflare owns CDN-to-compute and some database, and this deal gives Cloudflare a path toward an end-to-end stack agents already know how to use.

• 6

  Theo sees Lakebed and Void as the same category: his own unreleased project, Lakebed, is an AI-first cloud abstraction where npx lakebed new and npx lakebed deploy create a live app with auth, database, and sync, which he says is exactly why Cloudflare should now scare Vercel more than before.