NVIDIA CEO says new Blackwell Ultra GB300 AI platform is in full-scale mass production

NVIDIA CEO Jensen Huang made a quick flight and visit to Taiwan yesterday, meeting with key executives at TSMC and announcing that its Blackwell Ultra GB300 is in "full production" and "successfully ramping up".

Jensen's visit to Taiwan also saw the NVIDIA CEO give deep gratitude to the US government for approving export licenses for its H20 AI GPUs to China, and teased that its next-generation Rubin AI GPU architecture is its "most advanced" AI architecture, cooking in TSMC's ovens as you read this.

NVIDIA's next-gen Rubin chips include a dedicated CPU, GPU, scale-up NVLink Switch, and a new silicon photonics processor, which means the entire stack is getting a huge Rubin-infused upgrade. Rubin is a different beast to Blackwell and Hopper, as the changes to Rubin are from the ground up: HBM, the process node, design, and more

NVIDIA's new GB300 NVL72 AI server connects 72 x GB300 Blackwell Ultra AI GPUs and 36 Arm Neoverse-based NVIDIA Grace CPUs in a rack-scale design, acting as a single gigantic GPU built for test-time scaling. NVIDIA's new GB300 NVL72 has AI models accessing the platform's massive performance gains to explore different solutions to problems and break down complex requests into multiple steps, resulting in higher-quality responses.

NVIDIA's next-gen GB300 NVL72 AI server racks feature 288GB of HBM3E memory per GPU and offering even more compute power than GB200 NVL72 AI servers, which will handle even more complex reasoning models including OpenAI O1 and DeepSeek R1 models.