At the U.S.-Saudi Investment Forum in Washington, D.C., Nvidia CEO Jensen Huang, appearing alongside Elon Musk, firmly dismissed growing concerns of an AI investment bubble. Speaking on Wednesday, Huang argued that today’s explosive spending reflects a profound, irreversible transformation in computing—not speculative hype.
“Go back to first principles,” Huang urged. Moore’s Law, which drove exponential CPU performance for decades, has stalled. Demand for computing power, however, continues to soar. General-purpose processors can no longer scale economically, forcing the industry toward accelerated computing led by GPUs—a transition Nvidia pioneered over 20 years ago.
The evidence is stark: six years ago, CPUs powered 90% of the world’s top 500 supercomputers; today, that figure has collapsed below 15%. GPUs now dominate.
Huang emphasized that the largest computing workloads shifted to GPUs long before generative AI captured headlines. Banks, payment networks, e-commerce giants, and advertisers already spend hundreds of billions annually processing massive data frames and running recommender systems—the true “engine of the internet.” These foundational workloads, not ChatGPT-style models, drove the initial GPU surge.
Generative and agentic AI, he said, are merely the latest layer atop an already justified multi-trillion-dollar infrastructure buildout.
Despite Huang’s confidence, high-profile investors remain wary. Peter Thiel recently exited Nvidia entirely, and Michael Burry has compared current cloud and AI capex to past bubbles. Yet with U.S. tech giants projected to spend nearly $400 billion this year and global AI investment potentially reaching $5.2 trillion by 2030, Nvidia insists the demand is real, structural, and only beginning.