Jensen Huang Shares Insights on Artificial Intelligence and AGI

Nvidia CEO Jensen Huang made significant remarks regarding the future of artificial intelligence, token unit costs, the five-layer AI architecture, and the importance of general-purpose graphics processing units.

Artificial Intelligence and the Energy Analogy

Jensen Huang compared artificial intelligence to energy from 100 to 200 years ago, stating that it remains invisible and difficult to measure until it is used. He noted that today, artificial intelligence has a unit price in dollars per million tokens, showing a direct parallel to the cost per kilowatt-hour.

Five-Layer Artificial Intelligence Architecture

Huang described artificial intelligence as a five-layer cake, emphasizing that energy sits at the very bottom, followed by chips, data center infrastructure, models, and data with applications at the very top. He stated that the true value lies in the top layer, where companies add purpose, context, and access.

The Role of General-Purpose Graphics Processing Units

Stating that Nvidia's general-purpose graphics processing units are present in every cloud, data center, edge, and robotic system, Huang conveyed that these hardware components are software-upgradable. He noted that this architectural compatibility, spanning fourteen generations, provides a major competitive advantage for the company.

Views on Regulation and Security

Arguing that failing to adopt artificial intelligence and falling behind is the worst-case scenario, Huang stated that security is the responsibility of the builder. He expressed that regulations should target actual harm rather than hypothetical risks, and that rapid progress makes models more realistic.