GTC DC 2025 | NVIDIA CEO Renxun Huang’s keynote speech

GTC DC 2025 | NVIDIA CEO Renxun Huang’s keynote speech

In the keynote speech, NVIDIA founder and CEO Huang Renxun outlined the latest blueprint for the AI era. From large-scale GPU deployment and quantum technology breakthroughs, to AI factories, robotics technology, and autonomous driving.

NVIDIA’s pivot, from CPU to GPU accelerated computing: For decades, CPU performance has steadily improved like a clock – until the failure of Danard’s Law. NVIDIA’s solutions include parallel computing, GPU, and accelerated computing. Accelerated computing began with the NVIDIA CUDA-X library that runs through the entire technology stack. “This is indeed a treasure of our company,” Huang Renxun emphasized.

2

Quantum Leap – NVQLink: Huang Renxun announced the launch of NVIDIA NVQLink, a quantum GPU interconnect technology that enables real-time CUDA-Q calls from QPUs with latency as low as about 4 microseconds.

Accelerating scientific development: Huang Renxun announced that a national level laboratory is collaborating with NVIDIA to build a new supercomputer to promote the development of future science.

Entering the AI factory, ultimate collaborative design: Huang Renxun said, “AI is not a tool, AI itself is labor. For the first time, technology can truly take on work and help us improve productivity. This transition from tools to AI ‘workers’ is creating new computing models and bringing new job opportunities and industries along with it. ”The rise of AI factories is giving rise to new professions in fields such as AI engineering, robotics, quantum science, and digital operations. In addition, he also showcased the all-new NVIDIA BlueField-4 DPU.

3

Omniverse DSX officially released – Gigawatt level AI factory blueprint: Huang Renxun also introduced Omniverse DSX, which is a comprehensive blueprint for designing and operating 100 MW to several gigawatts (gigawatt level) AI factories.

NVIDIA Open Models, Data, Libraries: Huang Renxun explained that open source and open models drive innovation for startups, large enterprises, and researchers worldwide. This year alone, NVIDIA has contributed hundreds of open models and datasets. Subsequently, Huang Renxun highlighted the achievements of AI startups using NVIDIA technology, as well as the work from companies such as Google, Microsoft Azure, Oracle, ServiceNow, SAP, Synopsys, Cadence, CrowdStrike, and Palantir.

NVIDIA collaborates with leading global enterprises to create a digital twin platform: Physical AI is driving the process of reindustrialization, using robotics technology and intelligent systems to transform factories, logistics, and infrastructure. Huang Renxun emphasized in a video how his partners put it into practice. Huang Renxun also introduced Foxconn, Caterpillar Figure AI、 Jobs at companies such as Johnson&Johnson and Disney.

4

DRIVE Hyperion 10 helps the world’s leading enterprises to innovate the travel experience: Huang Renxun announced that Uber and NVIDIA are cooperating to build the infrastructure of autonomous travel, so as to realize the large-scale deployment of autonomous vehicle. In addition, Lucid、 Mercedes Benz and Stellantis are using the DRIVE AGX Hyperion platform to create L4 passenger cars.

 
© 版权声明
THE END
If you like it, please support it
点赞14 分享
comment 抢沙发

请登录后发表评论

    暂无评论内容