NVIDIA as we speak introduced on the Computex confence in Taiwan NVIDIA DGX Cloud Lepton — an AI platform with a compute market that connects builders constructing agentic and bodily AI functions with tens of hundreds of GPUs from a community of cloud suppliers, together with CoreWeave, Crusoe, Firmus, Foxconn, GMI Cloud, Lambda, Nebius, Nscale, Softbank Corp. and Yotta Information Providers.
The platforms will supply NVIDIA Blackwell and different NVIDIA structure GPUs on the DGX Cloud Lepton market.
Builders can faucet into GPU compute capability in particular areas for each on-demand and long-term computing, supporting strategic and sovereign AI operational necessities. Main cloud service suppliers and GPU marketplaces are anticipated to additionally take part within the DGX Cloud Lepton market.
“NVIDIA DGX Cloud Lepton connects our community of world GPU cloud suppliers with AI builders,” mentioned Jensen Huang, founder and CEO of NVIDIA. “Along with our NCPs, we’re constructing a planetary-scale AI manufacturing unit.”
NVIDIA mentioned DGX Cloud Lepton helps safe high-performance GPU assets by unifying entry to cloud AI providers and GPU capability throughout the NVIDIA compute ecosystem. The platform integrates with the NVIDIA software program stack, together with NVIDIA NIM and NeMo microservices, NVIDIA Blueprints and NVIDIA Cloud Features, to speed up and simplify the event and deployment of AI functions.
DGX Cloud Lepton offers administration software program for cloud platforms with features that ship real-time GPU well being diagnostics and automates root-cause evaluation, eliminating handbook operations and lowering downtime.
Options embrace:
- Improved productiveness and suppleness: Provides a unified expertise throughout growth, coaching and inference, serving to enhance productiveness. Builders can buy GPU capability straight from taking part cloud suppliers by way of {the marketplace} or carry their very own compute clusters, giving them larger flexibility and management.
- Frictionless deployment: Permits deployment of AI functions throughout multi-cloud and hybrid environments with minimal operational burden, utilizing built-in providers for inference, testing and coaching workloads.
- Agility and sovereignty: Offers builders fast entry to GPU assets in particular areas, enabling compliance with knowledge sovereignty laws and assembly low-latency necessities for delicate workloads.
- Predictable efficiency: Supplies taking part cloud suppliers enterprise-grade efficiency, reliability and safety, guaranteeing a constant consumer expertise.
A New Bar for AI Cloud Efficiency
NVIDIA as we speak additionally introduced NVIDIA Exemplar Clouds to assist NCPs improve safety, usability, efficiency and resiliency, utilizing NVIDIA’s experience, reference {hardware} and software program and operational instruments.
NVIDIA Exemplar Clouds faucet into NVIDIA DGX Cloud Benchmarking, a complete suite of instruments and recipes for optimizing workload efficiency on AI platforms and quantifying the connection between price and efficiency.
Yotta Information Providers is the primary NCP within the Asia-Pacific area to hitch the NVIDIA Exemplar Cloud initiative.
Builders can join for early entry to NVIDIA DGX Cloud Lepton.
