Skip to main content

Fluid compute is now the default for new projects

By Steven Van ·

Fluid compute reuses function instances to cut costs by up to 85% for high-concurrency workloads, and existing projects can opt in from the dashboard.

New Vercel projects now run on Fluid compute by default, as of April 23, 2025. Vercel says the change follows Fluid's general availability, its use in large-scale production apps, and all v0.dev deployments shipping with it enabled.

Fluid compute reuses existing function instances before starting new ones. Vercel says this cuts costs by up to 85% for high-concurrency workloads. Multiple invocations can share one instance on the Node.js and Python runtimes, and background work can continue after a response is sent using waitUntil.

Existing projects are not switched automatically. Fluid can be enabled from the project's Functions Settings in the dashboard, followed by a redeploy, or by setting "fluid": true in vercel.json for specific environments or deployments.

How multiple requests are processed in the fluid compute model with optimized concurrency.
How multiple requests are processed in Fluid compute with optimized concurrency
Vercel
Vercel
The platform for frontend developers — deploy, preview, and scale web apps and AI agents with zero config.
View Vercel →

Original source

Read Fluid compute is now the default for new projects on Creators Toolbox