NVIDIA has unveiled the Rubin CPX new generation of graphics processing unit (GPU) designed to handle vast amounts of AI data and enable smarter, more efficient applications for enterprises.
Expected to be available from end 2026, the new GPU lets AI systems process up to a million tokens — enough to analyse lengthy codebases or produce advanced generative videos.
Rubin CPX features advanced video processing and attention mechanisms that make long-context tasks possible without sacrificing speed or accuracy. Supported by NVIDIA’s enterprise software stack, these new GPUs promise to accelerate business innovation and keep enterprises at the forefront of AI development.
Working together with NVIDIA’s Vera CPUs in the Vera Rubin NVL144 CPX platform, it delivers eight exaflops of AI performance and 100 terabytes of fast memory in a single rack. This enables enterprises to run extremely demanding AI workloads, such as in-depth video analysis or intelligent software development assistance, with up to 7.5 times the performance of previous systems.
“Just as RTX revolutionised graphics and physical AI, Rubin CPX is the first CUDA GPU purpose-built for massive-context AI, where models reason across millions of tokens of knowledge at once,” said Jensen Huang, Founder and CEO of NVIDIA.
Enterprises benefit from efficient operations, with potential for significant returns. NVIDIA estimates up to US$5 billion in token revenue for every US$100 million invested in the platform.
Rubin CPX’s blend of scalability, rapid processing and advanced memory technology will enable enterprises deploying advanced AI solutions to achieve greater efficiencies and financial returns, effectively raising the bar for future AI infrastructure investments.
Early interest
Cursor, Runway and Magic are among those exploring Rubin CPX to enhance AI-driven code creation, generate cinematic content at scale, and enable AI agents that can handle extensive software projects autonomously.
“With NVIDIA Rubin CPX, Cursor will be able to deliver lightning-fast code generation and developer insights, transforming software creation. This will unlock new levels of productivity and empower users to ship ideas once out of reach,” said Michael Truell, CEO of Cursor, an AI-powered software company.
“Video generation is rapidly advancing toward longer context and more flexible, agent-driven creative workflows. We see Rubin CPX as a major leap in performance, supporting these demanding workloads to build more general, intelligent creative tools. This means creators — from independent artists to major studios — can gain unprecedented speed, realism and control in their work,” said Cristóbal Valenzuela, CEO of generative AI company Runway.
“With a 100-million-token context window, our models can see a codebase, years of interaction history, documentation and libraries in context without fine-tuning. This enables users to coach the agent at test time through conversation and access to their environments, bringing us closer to autonomous agentic experiences. Using a GPU like NVIDIA Rubin CPX greatly accelerates our compute workloads,” said Eric Steinberger, CEO of Magic, an AI research and product company.
Entelechy Asia 