Nvidia CEO Jensen Huang delivered a stark forecast at Stanford University, stating the next evolution of artificial intelligence will require a thousand times more computing power than current systems. Speaking in May 2026, Huang outlined that "agentic AI"-systems that can reason, plan, and act autonomously-will drive this exponential leap.
The demand is not a simple linear increase. Huang projected a hundred-fold increase in users interacting with these advanced models. The resulting energy consumption for AI computing is expected to climb to roughly 1,000 times current levels, forcing a reevaluation of global grid capacity and renewable energy investment.
The market is already feeling the pressure. Huang noted that Nvidia GPUs sold four to five years ago are seeing their resale prices rise, defying typical depreciation. This signals a deepening structural shortage in the semiconductor market, where demand is outpacing Nvidia's manufacturing capacity.
While Huang did not mention cryptocurrency, the impact is direct. GPU miners now compete directly with AI enterprises for scarce hardware. As AI scales, mining operations will face constrained supply and higher costs. Investors are watching secondary effects across the semiconductor supply chain, including high-bandwidth memory and data center cooling systems.