Nvidia Launches Global GPU Marketplace Through Rebranded Lepton Acquisition

Nvidia’s Lepton-Based GPU Marketplace Connects Developers to Global Network of Cloud Providers for AI Development

Three Key Facts

• Nvidia acquired Lepton AI in April 2025 and relaunched it as DGX Cloud Lepton – The Chinese startup specializing in GPU cloud services had raised $11 million in seed funding before the undisclosed acquisition
• DGX Cloud Lepton provides access to tens of thousands of GPUs – The marketplace connects developers globally to GPU resources from cloud providers including AWS, CoreWeave, Oracle, and Lambda through a unified platform
• Nvidia is constructing a 10,000+ GPU industrial AI cloud facility in Germany – The facility will serve European manufacturers including BMW, Mercedes-Benz, and Maserati for AI-driven manufacturing applications

Introduction

Nvidia has launched an ambitious initiative to create what executives describe as a “planet-scale AI factory” through its newly acquired and rebranded DGX Cloud Lepton platform. The company acquired Chinese startup Lepton AI in April 2025 and transformed it into a comprehensive GPU marketplace that addresses the global shortage of AI computing resources.

The platform represents a significant shift in how developers access high-performance computing power for artificial intelligence applications. Rather than competing directly with cloud hyperscalers, Nvidia has chosen to partner with them and smaller providers to aggregate GPU capacity into a single, accessible marketplace.

This strategic move comes as demand for AI computing resources continues to outpace supply, creating bottlenecks for organizations seeking to develop and deploy AI applications. The initiative positions Nvidia to capture value across the entire AI infrastructure stack while maintaining relationships with both large and emerging cloud providers.

Key Developments

DGX Cloud Lepton operates as a unified platform that aggregates GPU capacity from multiple cloud providers through a consistent software interface. Developers can access computing resources regardless of the underlying infrastructure location, simplifying the traditionally complex process of securing high-performance computing power.

The platform supports three primary workflows that address different stages of AI development. Dev Pods provide interactive development environments including Jupyter notebooks and VS Code for initial experimentation. Batch Jobs handle large-scale, non-interactive workloads such as model training that require sustained computing power.

Inference Endpoints enable organizations to deploy scalable, high-availability models for production use. The platform includes operational features such as real-time monitoring, auto-scaling capabilities, custom workspaces, and flexible security and data compliance options to meet enterprise requirements.

Market Impact

The marketplace model directly addresses the acute shortage of GPU resources that has constrained AI development across industries. By connecting developers to a broader network of providers beyond traditional hyperscalers, Nvidia democratizes access to high-performance computing infrastructure.

Major partnerships have already formed with global cloud providers including AWS, Microsoft Azure, Oracle, and specialized AI infrastructure companies like CoreWeave and Lambda. According to Forbes, the platform enables access through a global network that spans multiple continents and regulatory jurisdictions.

Oracle’s integration demonstrates the platform’s growing adoption among hyperscalers. The collaboration provides developers access to Oracle’s high-performance GPU clusters through DGX Cloud Lepton, with Oracle becoming one of the first major cloud providers to integrate with the marketplace.

Strategic Insights

Nvidia’s approach reflects a sophisticated understanding of evolving market dynamics as hyperscalers develop their own AI chips. By creating a platform that aggregates supply rather than competing directly, the company maintains relevance while hedge against potential revenue loss from traditional customers.

The initiative extends beyond simple resource aggregation to include comprehensive software integration. DGX Cloud Lepton incorporates Nvidia’s advanced software stack, including NIM, NeMo microservices, and Cloud Functions, creating a cohesive development environment that spans multiple providers.

Regional flexibility has become a crucial differentiator, particularly for regulated industries. The platform allows organizations to select GPU resources by geographic region to meet compliance requirements and data sovereignty needs that are increasingly important in finance, healthcare, and government applications.

Expert Opinions and Data

Industry leaders view the marketplace as a strategic response to changing market conditions. “Oracle has become the platform of choice for AI training and inferencing, and our work with NVIDIA boosts our ability to support customers running some of the world’s most demanding AI workloads,” said Karan Batta, senior vice president at Oracle Cloud Infrastructure.

The European expansion demonstrates Nvidia’s commitment to addressing sovereign AI requirements. The company is building a dedicated industrial AI cloud facility in Germany that will house over 10,000 GPUs and serve as a model for regional AI infrastructure development.

Nvidia CEO Jensen Huang articulates the broader vision behind the initiative: “In the era of AI, every manufacturer needs two factories: one for making things, and one for creating the intelligence that powers them.” This philosophy drives the company’s focus on enabling AI factories both in cloud environments and on-site installations.

The platform has attracted significant interest from venture capital firms and startups. Nvidia collaborates with European venture capital firms including Accel and Sofinnova Partners, offering up to $100,000 in GPU capacity credits and specialist support for eligible portfolio companies.

Early adopters span diverse industries and use cases. Organizations like Almawave utilize the infrastructure for multilingual generative AI model training, while healthcare companies like Cerebriu advance clinical MRI scan analysis through AI-accelerated tools integrated with the platform.

Summary

Nvidia’s DGX Cloud Lepton represents a fundamental shift in AI infrastructure strategy, moving from hardware sales to platform orchestration. The company has successfully transformed an acquired startup into a comprehensive marketplace that addresses critical supply constraints in the AI computing market.

The platform’s success depends on continued partnerships with cloud providers and adoption by developers seeking flexible, scalable access to GPU resources. Early indicators suggest strong market reception, with major hyperscalers integrating their services and enterprise customers beginning to leverage the unified platform for production workloads.

The initiative positions Nvidia to capture value across the AI development lifecycle while maintaining strategic relationships with both established cloud giants and emerging specialized providers. This approach creates multiple revenue streams and reduces dependence on traditional hardware sales channels as the AI infrastructure market continues to evolve.