Vercel Functions can now use Fluid compute on all plans. The new execution model lets a function instance handle several requests, rather than treating every invocation as a separate instance.
The launch includes Node.js and Python support, cold-start protection and a longer function lifecycle through background work after the response.
Why the execution model matters
- Concurrent requests can share an instance’s resources.
- Background tasks need a deliberate lifecycle.
- Resource use and invocation patterns affect the eventual bill.
Vercel’s published savings depend on the workload. Teams should compare their own traffic and resource settings rather than applying the headline reduction to every application. The deployment platform still has separate considerations for data services and other billable features.