Hugging Face has introduced ZeroGPU, a platform that enables developers to build and deploy AI applications without requiring dedicated GPU infrastructure. The service allows users to host up to 10 ZeroGPU Spaces with a $9 monthly subscription, providing 40 minutes of Blackwell GPU access per day. Developers can use any coding agent that supports the /goal command, such as Codex CLI or Claude Code, to automate the creation, deployment, and testing of AI apps.
The platform's key feature is its ability to attach GPUs only when needed, reducing costs and idle time. According to Hugging Face, the ZeroGPU platform is the cheapest serious compute on the internet for shipping demos to a wide audience. The platform uses the Gradio SDK, PyTorch 2.8+, and Python 3.10 or 3.12.
Developers can push code changes and receive a public URL instantly, with no need for Docker, Kubernetes, or GPU leases. The ZeroGPU platform also allows anonymous users to access 2 minutes of GPU time daily, while PRO users get 40 minutes. The platform's unique approach to GPU usage and its integration with Hugging Face's ecosystem make it a compelling option for developers looking to build and deploy AI applications efficiently.
Source: huggingface
