Nvidia Corp (NASDAQ:NVDA, ETR:NVD) has launched a new marketplace, DGX Cloud Lepton, designed to give AI developers easier access to high-performance computing power, as demand for GPUs continues to outstrip supply.
The platform acts as a global exchange where developers can source GPU capacity from a growing list of cloud providers, including CoreWeave, Foxconn, SoftBank and others.
It is aimed at easing the pressure on AI startups and enterprises struggling to secure the hardware needed to train and deploy large-scale models.
The marketplace supports both on-demand and longer-term use, allowing developers to rent compute power in specific regions, an increasingly important feature for companies managing data sovereignty and latency requirements.
Jensen Huang, Nvidia’s founder and chief executive, said DGX Cloud Lepton would connect a global network of GPU providers to create a “planetary-scale AI factory”.
The move comes as AI developers face growing complexity in managing workloads across different cloud services.
DGX Cloud Lepton aims to simplify that by integrating with Nvidia’s software stack, including tools for model training, testing, and deployment. It also provides cloud operators with real-time diagnostics and automated troubleshooting to reduce downtime and manual maintenance.
Alongside the launch, Nvidia introduced Exemplar Clouds, a framework to help cloud partners improve the performance and security of their AI services using Nvidia’s benchmarking tools and best practices.
Yotta Data Services has become the first participant in Asia Pacific to join the Exemplar Cloud initiative.
Early access to DGX Cloud Lepton is now open, with a broader rollout expected to follow.