Nvidia CEO Jensen Huang is betting big on artificial intelligence, projecting a multitrillion-dollar industry with what he calls 'very clear and good return on investment.'
Huang forecasts global AI infrastructure spending will reach between $3 trillion and $4 trillion by the end of the decade. During his CES 2025 keynote, he called AI agents a 'multi-trillion-dollar opportunity,' highlighting agentic AI-systems that plan, reason, and execute multi-step tasks autonomously.
Nvidia reports a $1 trillion order backlog for AI solutions through March 2026, with cumulative revenue from AI chips expected to hit at least $1 trillion by 2027. Robotics is also a key growth area, leveraging Nvidia’s Omniverse and Isaac platforms for simulation training.
The competition is intensifying. AMD’s MI300 series chips, Google’s custom TPUs, and Amazon’s Trainium are all vying for market share. Yet Nvidia’s moat remains its CUDA software ecosystem, built over a decade.
Huang’s emphasis on ROI addresses concerns over massive AI spending without proportional productivity gains. While Google, Amazon, and Microsoft develop in-house chips, none have matched Nvidia’s performance.