Jensen HuangFounder, President and Chief Executive Officer, NVIDIA Corporation (NVDA)
9 recorded appearances
With Garry Tan: NVIDIA's near-death moments, starting on the wrong algorithm and relearning the domain from textbooks, first-principles culture, systems thinking as the new coding, agentic AI inside NVIDIA, and the physical-AI "ChatGPT moment" for robotics.
With Mike Allen: hard pushback on extreme AI-risk narratives and job-loss predictions, China policy and Chinese models, what Wall Street misreads, and timelines for useful humanoid robots.
Taiwan ecosystem focus: agentic-AI progress read off the GitHub commit explosion, Vera Rubin production status, Cosmos world models, and the physical-AI robotics platforms.
The cleanest framing of the shift from retrieval to generative/agentic computing: AI factories as the new dynamos, the five-layer stack from energy to applications, and why he thinks AI creates more jobs than it destroys.
Argues Nvidia's durable advantage is the bottlenecked advanced-chip supply chain rather than the GPU alone, defends selling AI chips into China, and explains why Nvidia will not become a hyperscaler itself.
Recorded straight after the GTC 2026 keynote: the growing CPU business and the Groq acquisition, where scarcity sits in the AI stack, and the China question.
The deepest long-form dive: extreme co-design and rack-scale systems, scaling laws including agentic, power and memory bottlenecks, the supply chain, China and Taiwan, the moat, data centres in space, and his AGI timeline.
The announcement-heavy one: the inference inflection, AI factories, the Vera Rubin architecture, agentic-AI platforms, physical AI and robotics, and open models.
The loosest and most personal: origin stories and near-deaths, energy as the real bottleneck, AI-risk scepticism, manufacturing and energy policy, and some colourful predictions.