Platform SettingsFluid Compute
Fluid Compute
One toggle that can cut function costs by 80%+.
TL;DR
Enable Fluid Compute in your Vercel dashboard. It reuses warm instances across requests, reducing cold starts and potentially cutting costs by 80%+. Zero code changes required.
Traditional vs Fluid
The magic: I/O wait time stops counting. You only pay for actual JavaScript execution.
If your function spends 90% of time waiting for external services, Fluid Compute reduces costs by ~90%.
Enable It
Go to project settings
Dashboard → Your Project → Settings
Find Functions tab
Left sidebar → Functions
Enable Fluid Compute
Toggle it on. Requires Standard CPU (1x).
Redeploy
Push any commit to apply.
Expected Savings
| Your App Type | Savings |
|---|---|
| LLM integrations (OpenAI, etc.) | 85-95% |
| Database-heavy dashboards | 70-85% |
| External API aggregators | 75-90% |
| Payment processing | 65-80% |
| PDF generation | 40-60% |
| Pure computation | 15-30% |
The pattern: more I/O waiting = more savings.
Next: Build Settings for deployment cost reduction.