Nvidia CEO set to unveil new AI chips and software

During a keynote speech expected to draw more than 18,000 attendees at a hockey arena, Huang will outline Nvidia’s strategy for adapting to the rapidly evolving artificial intelligence landscape, News.Az reports, citing Reuters.

The company, currently the world’s most valuable publicly listed firm with a market capitalization exceeding $4.3 trillion, is expected to introduce a next-generation AI chip named Feynman, honoring American physicist Richard Feynman.

The four-day conference will also feature discussions on data centers, Nvidia’s CUDA chip programming software, AI agents, and physical AI applications such as robotics.

Another key topic is likely to be Groq, a chip startup whose technology Nvidia licensed for $17 billion in December. Groq specializes in fast and cost-efficient “inference” computing, where AI models apply learned data to answer questions or generate predictions in real time.

In recent years, technology companies have spent hundreds of billions of dollars on chips to train AI models. However, firms such as OpenAI, Anthropic and Meta Platforms are now shifting focus toward serving hundreds of millions of users who rely on these AI systems.

As demand grows for inference computing, Nvidia faces increasing competition, including from some of its own customers developing custom chips. Analysts expect the company to strengthen its position against rivals seeking to regain market share lost to Nvidia.

Despite the rising competition, Nvidia remains a central player in the global AI ecosystem. Countries including Saudi Arabia are building customized AI systems using its chips, and the company remains one of the few major U.S. tech firms continuing to release open-source AI software, an area increasingly tied to technological rivalry between the United States and China.