Nvidia CEO Jensen Huang is set to outline the company’s upcoming hardware and software strategy at its annual developer conference in San Jose on Monday.
During a keynote address expected to take place in a hockey arena with capacity for more than 18,000 attendees, Huang will present Nvidia’s roadmap as the AI industry evolves rapidly and demand for advanced computing power continues to surge.
Nvidia — currently the world’s most valuable listed company with a market capitalisation exceeding $4.3 trillion — is widely expected to unveil details of a next-generation AI chip known as Feynman, named after American physicist Richard Feynman. The announcement is likely to be one of the major highlights of the four-day conference.
Huang is also expected to discuss Nvidia’s broader AI ecosystem, including data centre technology, its widely used chip programming platform CUDA, digital assistants known as AI agents and developments in physical AI such as robotics.
This year’s event carries heightened importance as investors look for reassurance that Nvidia’s strategy of reinvesting large portions of its profits into the AI ecosystem is delivering long-term returns.
Another key topic could be Groq, a chip startup whose technology Nvidia licensed for $17bn in December. Groq focuses on high-speed and cost-efficient inference computing, where AI models apply previously learned knowledge to generate answers or predictions in real time.
The shift toward inference computing is gaining momentum as companies such as OpenAI, Anthropic and Meta Platforms increasingly focus on serving hundreds of millions of users who interact with AI systems.
Nvidia faces stronger competition in the inference-chip market than in AI training chips, and analysts expect the company to outline strategies to defend its dominance against rivals attempting to reclaim market share.
Investors are also looking for further explanation behind Nvidia’s recent $2bn investments in Lumentum and Coherent, both of which develop laser technologies used to transmit data between chips via beams of light — a critical component for future high-performance AI systems.
