Vercel Functions Now Run Up to 30 Minutes
Serverless just got a longer leash. Vercel has raised the Vercel Functions 30-minute execution limit for Pro and Enterprise teams running Node.js and Python runtimes. That is more than double the previous ceiling of 800 seconds. Consequently, this changes what teams can realistically build without leaving Vercel’s platform.
The update directly targets workloads that simply need more time. Long LLM reasoning chains, multi-minute AI response streams, document processing, OCR, data extraction, web scraping, and complex Workflow or Queue steps are all now viable inside a standard Vercel Function. Previously, these tasks required external infrastructure or careful orchestration to stay within the time cap.
Cost efficiency was also a clear design consideration. Vercel’s Fluid Compute model bills only during active CPU execution. So when a Function waits on an external call, an AI model response, a database query, or a third-party API the billing clock pauses. Therefore, a 30-minute function spending most of its time waiting on model inference does not bill like a 30-minute continuously running process. That is a meaningful difference for AI-heavy teams watching infrastructure costs.
Opting in is straightforward. For Next.js App Router, teams set maxDuration to 1800 directly in the route file. For other frameworks, the same value goes into vercel.json under the specific function path. Moreover, Vercel notes that durations above 800 seconds are currently in beta and require Fluid Compute to be enabled first.
Support for additional runtimes beyond Node.js and Python is coming soon. For teams building AI-heavy applications or backend pipelines constrained by execution time limits, this update removes one of the most practical obstacles to staying on a serverless-first architecture. Overall, the Vercel Functions 30-minute execution limit marks a clear shift in what serverless platforms now consider standard.
