← All posts

Khoury College of Computer Sciences

CS 4550 / CS 5610 · Web Development · Fall 2026

Vercel Fluid: one compute layer for functions, sandboxes, and builds

  • Next.js
  • AI
  • Course

Instructor-curated digest, not original reporting. Each post paraphrases a public article and links to the original source.

On September 1, 2026, Luke Phillips-Sheard published Compute that takes any shape on the Vercel blog. Fluid is Vercel's unified compute layer: it assembles the machine a workload needs, swaps configuration on the fly, and absorbs burst capacity. Builds ran on it first, then sandboxes, and now functions; the post says if you have shipped on Vercel you have already been on Fluid.

The post cites more than 15 million builds a day, 25 million sandboxes a week, and a trillion requests a month. Building blocks include Hive, which provisions isolated machines; Fluid images and VHS snapshots, which resume custom environments rather than cold-booting; and Vercel Drives, portable storage that outlives the machine (sandboxes today, private beta). Many requests share one instance, and Active CPU pricing means you pay for CPU while code runs, not while waiting on a database or model. The write-up is framed for agents that need fast, isolated machines with their own environment and state that outlives compute.

Fluid is the compute under the same Vercel deploy path you already use in Chapter 1 (and the Node server work in Chapter 5).

Read original: Compute that takes any shape — Vercel

Related in the book: Chapter 1 · Chapter 5