Jensen Huang

CEO of Nvidia, referenced in Chip War for two moves that shaped the modern compute landscape. In 2006, recognizing that high-speed parallel computation had uses beyond graphics, Nvidia released CUDA — software that let GPUs be programmed in a standard language without any reference to graphics at all. Huang spent lavishly on this software effort (at least $10 billion by a 2017 company estimate) to let any programmer, not just graphics experts, work with Nvidia’s chips. That bet became the foundation of Nvidia’s later dominance in AI compute.

He is also cited in Chip War among the industry luminaries — alongside former Stanford president and Alphabet chairman John Hennessy — who have declared Moore’s Law dead, on the view that the laws of physics will eventually make it impossible to keep shrinking transistors, and that the tools needed (notably $100M+ EUV lithography machines) are already staggeringly expensive.

Context: Jensen Huang co-founded Nvidia in 1993 and has led it since. The CUDA platform he championed became the dominant software substrate for GPU-accelerated computing and, later, the deep-learning boom, making Nvidia one of the most valuable companies in the world.

Where this appears

• Chip War — credited with the CUDA bet that opened GPUs to general-purpose computing, and quoted as one of the executives declaring Moore’s Law dead.
• Who’s Afraid of Chinese Models? — his framing of Nvidia GPUs as “token factories” (tokens-per-second, cost-per-token metrics), engaged by Ben Thompson’s argument that reasoning-model token inefficiency complicates that framing.
• AI Is Oil, Not God — signatory of the July 2026 open-weights letter.
• Write Things Down — declared on X that AGI has arrived in the form of OpenAI’s “Astra” model; Ben Thompson respectfully dissents, arguing continuous weight updates should be the AGI threshold.