General availability for Flash, its tool for running serverless GPU and CPU workloads in pure Python without requiring Docker. The release is positioned as production-ready infrastructure for developers who want faster deployment loops.
Dockerless workflows can reduce friction for teams experimenting with AI workloads, especially when they need to move quickly from notebook-style code to repeatable jobs. Serverless GPU access also helps teams avoid managing idle infrastructure.
The launch broadens RunPod's infrastructure stack at a time when AI developers are looking for simpler ways to deploy custom inference and batch workloads.