According to Nvidia CEO Huang He-juen's recent interview, the global economy is undergoing an infrastructure wave potentially exceeding the Industrial Revolution, with AI agents at its center. Huang stressed that AI has transitioned from generative capabilities—such as text-to-text and text-to-image conversion—to an "agentic" stage where systems can reason, execute tasks, and complete concrete work. This shift represents the true turning point for AI's commercial value, as markets will pay for "completed work" rather than merely demonstrated intelligence.
Huang further outlined a five-layer AI industry structure: energy (nuclear, wind, solar, hydrogen); computing hardware (chips, servers, network equipment); infrastructure (land, power, data centers); model development (large language model creators like OpenAI); and applications (finance, law, logistics, manufacturing). He emphasized that AI learning extends beyond language to any domain with stable structure and predictable patterns, suggesting real-world and life sciences applications could represent a market far larger than currently recognized.