Uploaded April 2026 | Updated September 2026, 2 weeks ago
This talk outlines how we design and operate a global GPU research cluster that supports AI and visual computing work for our console platform and game development studios. We use NVIDIA GPUs to train advanced models that run on PlayStation hardware to improve rendering, gameplay systems, and overall player experience. Running shared GPU infrastructure at scale brings challenges: securing data center space, long procurement cycles, multi-tenant Kubernetes management, and noisy-neighbor pressures as research needs shift. We will share lessons in cluster design, planning, security, and operations that help studios innovate and push the console experience forward.
- Get started with NVIDIA tools and resources for game developers: developer.nvidia.com/industries/game-development
This talk outlines how we design and operate a global GPU research cluster that supports AI and visual computing work for our console platform and game development studios. We use NVIDIA GPUs to train advanced models that run on PlayStation hardware to improve rendering, gameplay systems, and overall player experience. Running shared GPU infrastructure at scale brings challenges: securing data center space, long procurement cycles, multi-tenant Kubernetes management, and noisy-neighbor pressures as research needs shift. We will share lessons in cluster design, planning, security, and operations that help studios innovate and push the console experience forward.
- Get started with NVIDIA tools and resources for game developers: developer.nvidia.com/industries/game-development
