Jensen Huang says AI agents could drive a 5-10x computing boom: "100 billion agents and billions of robots"

Nvidia CEO Jensen Huang recently projected a massive shift in the computing industry. During a recent interview, Huang predicted that the semiconductor sector needs to grow five to tenfold over the next decade to accommodate a new wave of autonomous AI agents and physical robots. He argued that the current infrastructure is insufficient for a future where computers primarily talk to other computers rather than people.

Huang estimates that the world will eventually see 100 billion AI agents and billions of robots operating around the clock. This shift changes the requirements for data centers and memory hardware. As these autonomous background loops generate the bulk of traffic, backend systems must move beyond traditional constraints like rate limiting and session memory that were built for human users.

To support this growth, Nvidia has entered a partnership with SK Group valued at over $500 billion. This agreement involves purchasing next-generation High-Bandwidth Memory (HBM) and co-developing custom roadmaps for HBM4 specifically for agentic AI workloads. SK Telecom plans to build a 2-gigawatt AI data center in South Korea using Nvidia’s Vera Rubin architecture, with operations set to begin in 2027.

Nvidia is also investing $1 billion into Naver Corp to increase data center capacity. Huang noted that South Korea occupies a critical role in the global supply chain, calling the current period a golden age for the country’s industrial capabilities. This proactive investment strategy mirrors Nvidia’s broader approach of securing advanced packaging, power, and memory capacity years before demand spikes.

Ultimately, Huang suggests that the industry must stop planning for a world where computers serve humans and start preparing for one where machine-to-machine traffic dominates. Whether the physical infrastructure can keep pace with this demand remains the primary challenge for the coming decade.