NVIDIA announced today at Computex Condence at NVIDIA DGX Cloud Lepton, Taiwan, that it is an AI platform with a computing market that will enable developers to establish proxy and physical AI applications with networks including cloud providers, including tens of thousands of GPUs, including cloud provider network data services.
These platforms will be available on the DGX Cloud Lepton Marketplace for Nvidia Blackwell and other NVIDIA architecture GPUs.
Developers can leverage GPU computing power in specific areas of on-demand and long-term computing to support strategic and sovereign AI operational requirements. Leading cloud service providers and GPU markets are also expected to participate in the DGX Cloud Lepton Marketplace.
“NVIDIA DGX Cloud Lepton connects our global network of GPU cloud providers with AI developers,” said Jensen Huang, founder and CEO of NVIDIA. “With our NCP, we are building a planet-scale AI factory.”
NVIDIA says DGX Cloud Lepton helps ensure high-performance GPU resources by unifying access to cloud AI services and access to the NVIDIA COMPECTUTE ecosystem. The platform integrates with the NVIDIA software stack, including NVIDIA NIM and NEMO microservices, NVIDIA Blueprints and NVIDIA Cloud capabilities to accelerate and simplify the development and deployment of AI applications.
DGX Cloud Lepton provides management software for cloud platforms that provide real-time GPU health diagnosis and automate root cause analysis, eliminating manual operations and reducing downtime.
Functions include:
- Improve productivity and flexibility: Provide a unified experience in development, training, and reasoning to increase productivity. Developers can purchase GPU capacity directly from participating cloud providers through the market, or bring their own compute clusters, giving them more flexibility and control.
- Friction-free deployment: Use integrated services to reason, test and train workloads to deploy AI applications in multi-cloud and hybrid environments with minimal operational burden.
- Agility and Sovereignty: Allows developers to quickly access GPU resources in specific regions, complying with data sovereignty regulations and meeting low latency requirements for sensitive workloads.
- Predictable performance: Provide enterprise-grade performance, reliability, and security to participating cloud providers to ensure a consistent user experience.
New standards for AI cloud performance
NVIDIA also today announced the NVIDIA Model Cloud to help NCPS use NVIDIA’s expertise, reference hardware and software and operational tools to help enhance security, availability, performance and resilience.
NVIDIA Sample Cloud introduces NVIDIA DGX Cloud Bench Markess, a comprehensive set of tools and recipes to optimize workload performance on AI platforms and quantify the relationship between cost and performance.
Yotta Data Services is the first NCP in the Asia-Pacific region to join the NVIDIA Model Cloud Program.
Developers can register early to enter NVIDIA DGX Cloud Lepton.
