“OpenAI is very likely going to be the world’s next multitrillion-dollar hyperscale company,” Huang said.
That bold prediction comes at a moment when even AI’s loudest evangelists are warning of overvaluation and overbuilding. Altman himself has cautioned that too much money is flooding into unproven AI ventures, while Zuckerberg has compared today’s infrastructure frenzy to past bubbles. Yet Huang insists the skeptics are missing the deeper forces reshaping the economy. In his telling, the story comes down to basic physics, not hype.“General-purpose computing is over,” Huang said, describing what he sees as a generational shift in how all industries will run. “The future is accelerated computing and AI.”
He outlined what he calls the “three scaling laws” of AI—pretraining, post-training, and inference—each of which exponentially increases demand for compute. While training workloads have already been well-documented, Huang stressed that inference—the real-time reasoning that underpins everything from chatbots to recommendation algorithms—is only just beginning.
“The longer you think, the better the answer you get—and thinking requires more compute,” he explained.
That framing matters because inference is where AI collides with day-to-day usage. Training runs happen in bursts, but inference happens constantly: Every chatbot prompt, every AI video render, every background algorithmic tweak consumes processing power. If Huang is right, that relentless demand means AI won’t follow the boom-and-bust cycles of earlier technologies but will instead drive a compounding need, one that will also boost Nvidia.
