NEWARK, NJ – Runpod, the AI developer cloud, today announced the general availability of Runpod Flash, an open source Python SDK that removes the infrastructure between writing AI code and running it in production. With Flash, developers go from a local Python workload to a live, auto-scaling repository in minutes, with no containers to build, no images to manage, and no infrastructure to configure. Flash is now available on PyPI and GitHub under the MIT license.
How does this work
Flash supports two usage patterns. Queue-based processing handles batch and synchronized workloads. Load-balanced endpoints provide real-time traffic inference. Developers specify their computing needs and dependencies in Python, and Flash handles provisioning, scaling, and infrastructure management automatically.
Endpoints automatically scale from zero to a fixed size based on demand, then scale back when idle. Flash also includes a command-line interface for environment development, testing, and production deployment, giving developers a complete workflow from testing to deployment.
Beyond standalone endpoints, Flash Apps supports multi-point applications for production architectures that require different computing configurations to work together. Developers can prototype on Runpod Pods, package their logic with Flash, run on Serverless, and scale to production without switching providers. Flash Apps allow developers to combine multiple endpoints with different computing settings into a single usable service. An agent orchestration layer can run on one type of computer while the underlying conceptual model runs on another, all managed and scaled as a single unit. Combined with Runpod Serverless’s scale-to-zero economics, Flash becomes the natural computing backbone for agent systems that need to drive models on demand without paying for redundant infrastructure.
Why Runpod builds Flash
“We’ve built one of the largest serverless platforms in the industry, and Flash makes it very quick to get on board.” said Zhen Lu, Runpod’s CEO and co-founder. “Python’s native workflow becomes a live endpoint, scaled automatically in minutes, at the same billing per second and economies of scale to zero that our developers are already using. Flash is what continuous improvement looks like in mobile AI.”
“And we’re seeing a shift in the way AI applications are being built. Agents don’t fit neatly into one container or one storage location. They need to call on different models, route between different types of computing, and scale on demand. Flash and Runpod Serverless were designed for exactly that kind of work.”
Inference is the next phase of AI infrastructure
The AI infrastructure is changing. The first wave of investment in this industry was dominated by training: building base models that required massive, continuous computation. The next wave is conceptualization, where those models are put to work in production systems that serve real users. Virtual workloads now represent the fastest growing segment of AI cloud spending, and the tool requirements are very different: dynamic demand, time sensitivity, cost pressure at scale, and the need to deploy and iterate quickly.
Runpod has emerged as a major workload targeting platform. Over 750,000 developers use Runpod to build and deploy AI, with 37,000 serverless endpoints created in March 2026 alone and over 2,000 developers creating new endpoints every week. The teams at Glam Labs, CivitaI, and Zillow are using a productivity approach on the platform. The company has reached $120M in annual recurring revenue.
Flash accelerates this momentum by removing the last major point of contention in the deployment workflow. Rather than spending time on container configuration and registry management, developers can focus on the application concept and get to production faster.
Runpod’s position in AI infrastructure
The cloud AI market has grown past $7 billion with more than 200 providers, but developers still face a tough trade. Hyperscalers provide scale but come with complex toolchains, lock-in, and high costs. Neoclouds requires business contracts and minimum obligations. Point solutions handle a single task well but force developers to redesign as their needs change.
Runpod bridges the gap between these options: self-service access, native developer experience, full lifecycle coverage from testing through production, at an affordable cost. Flash extends that space by making the deployment experience similar to the simplicity of the rest of the platform.
