New York, August 19, 2025 – Lightning AI, the company best known for creating PyTorch Lightning and building the infrastructure layer for AI development, has announced the launch of its Multi-Cloud GPU Marketplace. The platform is the first of its kind to give AI teams unified access to both hyperscale cloud providers and a new class of specialized compute platforms called NeoClouds.
The marketplace promises to cut AI infrastructure costs by up to 70%, while eliminating long-standing tradeoffs between flexibility, speed, and scale.
A Unified Platform for GPUs Across Clouds
For years, AI developers have faced a difficult choice: stick with established hyperscalers that offer robust ML platforms, or turn to emerging providers and take on the burden of building infrastructure from scratch. Lightning AI’s new marketplace solves this by offering:
- On-demand and reserved GPU access across multiple clouds
- A single intuitive interface trusted by over 300,000 developers and Fortune 500 enterprises
- The ability to optimize for cost, performance, or region without vendor lock-in
“AI teams shouldn’t have to rebuild their entire stack every time they change providers,” said Luca Antiga, CTO of Lightning AI. “With Lightning AI, they don’t. They get access to the GPUs they want, wherever they want them, with their highly-specialized stacks, without slowing down.”
Choice Without Complexity
The marketplace supports both on-demand GPUs and large reserved clusters, with orchestration options including SLURM, Kubernetes, or Lightning’s own next-gen AI orchestrator. This flexibility lets teams bring their preferred tools and stacks with zero workflow disruption.
“Infrastructure costs, complexity, and lock-in have become obstacles as companies double down on AI,” said Luca Antiga, CTO of Lightning AI. “With Lightning AI, teams no longer need to rebuild their stack every time they switch providers.”
Partner integrations include:
- Lambda: Offering seamless access to high-performance on-demand compute|
- Nebius: Enterprise-grade AI infrastructure with zero DevOps overhead
- Voltage Park: Large-scale “AI factory” compute integrated directly into Lightning AI workflows
- Nscale: Secure, sovereign infrastructure for European enterprises
Key Benefits of Lightning AI’s Marketplace
- Run AI workloads across multiple clouds with one interface
- Reserve or scale on-demand depending on project needs
- Avoid vendor lock-in with a portable, flexible platform
- Eliminate DevOps overhead with fully managed orchestration options
- Scale training, fine-tuning, and inference without rework
About Lightning AI
Lightning AI is the company behind PyTorch Lightning, one of the world’s most popular deep learning frameworks. Its flagship products include:
- Lightning AI Studio – streamlines AI experimentation and development
- Lightning AI Hub – enables seamless deployment of scalable, enterprise-ready AI systems
By combining cutting-edge infrastructure with developer-first tools, Lightning AI empowers enterprises and research teams to accelerate their AI initiatives with confidence.
