What is Rubin
Rubin is NVIDIA's new-generation AI chip platform, comprising six chips: Vera CPU, Rubin GPU, NVLink 6 switch, ConnectX-9 network card, BlueField-4 DPU, and Spectrum-6 Ethernet switch. Through extreme co-design, the chips treat the data center as a single computing unit, enabling efficient, scalable, and secure AI computing. The Rubin platform supports large-scale model training, reduces inference costs by 10x, and improves training efficiency by several times. It is suitable for scenarios such as autonomous driving and industrial manufacturing, driving AI from the laboratory to industrial-grade applications.
Main Features of Rubin
- Efficient Computing: The Rubin platform supports large-scale model training and real-time inference, significantly improving the efficiency of AI tasks, reducing inference costs, and meeting the needs of complex AI applications.
- Co-design: By integrating GPU, CPU, networking, security, and other components into a unified design, the Rubin platform achieves data center-level overall optimization, improving the performance and efficiency of large-scale deployments.
- Networking and Scaling: With NVLink 6 and Spectrum-X Ethernet technology, the Rubin platform provides high-speed communication and large-scale scaling capabilities, supporting flexible deployment and efficient operation of AI systems.
- Security and Reliability: The Rubin platform features full-stack security capabilities, supports zero-downtime maintenance and dynamic traffic redirection, ensuring high reliability and security during the operation and maintenance of AI factories.
- Energy Efficiency: Through intelligent power management and liquid cooling technology, the Rubin platform optimizes energy utilization, reduces data center energy consumption, and improves overall energy efficiency.
- Development and Deployment: The Rubin platform supports software frameworks such as CUDA and CUDA-X, providing a mature ecosystem that helps developers quickly deploy and scale AI applications.
Rubin Official Website Address
The Rubin platform is expected to launch in the second half of 2026.NVIDIA's hardware is still in the launch phase. According to NVIDIA, AWS, Google Cloud, Microsoft, and OCI will all become the first public cloud platforms to deploy Vera Rubin instances in 2026.
Application Scenarios of Rubin
-
Autonomous Driving: The platform supports the development and deployment of autonomous vehicles, providing strong support for autonomous driving systems through efficient computing power and real-time inference capabilities, driving the development of autonomous driving technology.
-
Industrial Manufacturing: The platform can integrate with industrial software to achieve full lifecycle management from chip design to production operations, enhancing the intelligence level of industrial manufacturing and driving the new industrial revolution.
-
Robotics: The platform supports intelligent robot development, providing efficient data processing and inference capabilities, facilitating the application of robots in different scenarios, and enhancing the intelligence level of robotic systems.
-
Data Centers and Cloud Computing: The platform provides efficient, scalable AI computing power for data centers and cloud computing environments, supporting the training and inference of large-scale AI models and meeting the needs of cloud service providers.
-
Scientific Research: The platform supports high-performance computing, accelerating complex computational tasks in scientific research, such as climate simulation and biomedical research, driving scientific discovery and technological innovation.