Nvidia has rebranded Lepton AI as DGX Cloud Lepton and reintroduced it in June 2025. As stated by Nvidia, the service offers a unified AI platform and compute marketplace that links developers to tens of thousands of GPUs from a global network of cloud providers.

How Does DGX Cloud Lepton Work

DGX Cloud Lepton functions as a unified AI platform and marketplace, bridging the gap between developers and GPU resources worldwide. It combines GPU capacity from cloud providers like AWS, CoreWeave, and Lambda through a consistent software interface. This allows developers to access GPU computing via a centralized interface, no matter where the cluster is located.

While utilizing the underlying GPU compute, Nvidia provides a uniform software platform powered by NIM, Nemo, Blueprints, and Cloud Functions. Regardless of the cloud infrastructure used, developers can expect the same software stack for running their AI workflows.

DGX Cloud Lepton supports three main workflows:

Dev Pods: Interactive development environments (e.g., Jupyter notebooks, SSH, VS Code) for prototyping and experimentation.

Batch Jobs: Large-scale, non-interactive workloads (e.g., model training, data preprocessing) that can be spread across multiple nodes, with real-time monitoring and detailed metrics.

Inference Endpoints: Deploy and manage models (base, fine-tuned, or custom) as scalable, high-availability endpoints, supporting both NVIDIA NIM and custom containers.

In addition, DGX Cloud Lepton provides operational features such as real-time monitoring and observability, on-demand auto-scaling, custom workspaces, security measures, and compliance tools. Developers have the option to select their preferred region to ensure data locality and meet data sovereignty requirements.

DGX Lepton’s Expanding Network

Nvidia has formed partnerships with major global cloud and infrastructure providers. Andromeda, AWS, CoreWeave, Foxconn, Hugging Face, Lambda, Microsoft Azure, Mistral AI, Together AI, and Yotta are among the listed partners for DGX Cloud Lepton.

At the recent GTC event in Paris, Nvidia revealed collaborations with leading European cloud providers to help local developers fulfill data sovereignty needs. The company also announced a partnership with Hugging Face to offer training clusters as a service.

Nvidia is teaming up with European venture capital firms, including Accel, Elaia, Partech, and Sofinnova Partners, to provide up to $100,000 in GPU capacity credits and support from NVIDIA specialists for eligible portfolio firms through DGX Cloud Lepton.

Although pricing varies depending on the cloud provider, the service is currently in preview. Developers can register at http://m.miracleart.cn/link/512844c25cd1c9af079ba08f868eb637 to request early access to Lepton.

With DGX Cloud Lepton, Nvidia aims to make GPU computing available to developers globally. Rather than launching its own cloud platform to compete with hyperscalers, Nvidia has opted to collaborate with them to deliver aggregated compute resources to developers.

The above is the detailed content of Nvidia Wants To Build A Planet-Scale AI Factory With DGX Cloud Lepton. For more information, please follow other related articles on the PHP Chinese website!