NVIDIA has once again demonstrated its leadership in artificial intelligence and high-performance computing. The company announced the technical details of its Vera Rubin NVL72 processor and the next-generation Rubin GPU architecture. This new architecture promises up to a 10x performance increase in AI workloads compared to the previous Blackwell architecture.

The Rubin architecture specifically aims to accelerate the shipment of hundreds of thousands of independent Grace servers by increasing CPU density in data centers. This strategy will enable AI models to operate more efficiently and be deployed at scale. According to NVIDIA's statements, the Olympus cores of the Vera processor redefine speed limits for dedicated AI architecture.

Key Features and Innovations:

• 10x Performance Increase: A significant leap in AI performance compared to the Blackwell architecture.
• Vera Rubin NVL72 Processor: A powerful processor designed for next-generation AI workloads.
• Enhanced Data Center Integration: Architecture that increases CPU density, providing a more balanced and efficient operating environment in data centers.
• Olympus Cores: Cores focused on dedicated AI architecture, setting new standards in speed and efficiency.

These developments have the potential to accelerate the evolution of AI-based applications and systems, opening new doors in the industry. NVIDIA continues to shape the future of AI infrastructure with its Rubin architecture.