# Vercel — Fluid compute: Evolving serverless for AI workloads

- Company: Vercel (vercel.com)
- Announced: 2025-05-30T13:00:00+00:00
- Category: capability-change
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://vercel.com/blog/fluid-compute-evolving-serverless-for-ai-workloads
- Record: https://forck.live/items/20869-fluid-compute-evolving-serverless-for-ai-workloads
- Subject: v0 / AI SDK / AI Gateway

Vercel introduced Fluid compute, a new compute model designed for AI workloads that reuses existing serverless resources before scaling new ones. Fluid reduces total compute consumption by prioritizing existing resources, scaling inside a single instance, and avoiding cold starts. It is now the default for new projects and can be enabled for existing projects.

## Evidence

Verbatim from https://vercel.com/blog/fluid-compute-evolving-serverless-for-ai-workloads:

> Instead of launching a new function for every request, it intelligently reuses available capacity, ensuring that compute isn't sitting idle.

---

Record: https://forck.live/items/20869-fluid-compute-evolving-serverless-for-ai-workloads
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md