Nvidia CEO Jensen Huang told an audience in San Jose, California, that the AI boom’s next phase will be driven by a shift from building models to running them—an update he described as “inference inflection.” Speaking at the arena event on Monday, Huang framed the industry’s current momentum as an early stage of broader computing change, and he said Nvidia expects strong demand to persist through a growing orders pipeline for its chips.
Huang, 63, spent more than two hours moving across the stage as he outlined how Nvidia’s processors became central to AI workloads. He also discussed how he believes the AI buildup remains “in its infancy,” emphasizing that the industry is at “the beginning of a new platform change,” a message he compared to earlier transitions such as the PC and internet revolutions.
To underscore his view of demand, Huang predicted that Nvidia will be grappling with a $1 trillion backlog in chip orders by the end of the year, which he said would double his estimate from a year ago. He tied the company’s role to the practical mechanics of AI, saying inference chips let systems use what they have learned to produce responses—such as writing documents or creating images—more efficiently than processors used earlier while large language models were being built.
Huang’s remarks also highlighted Nvidia’s recent financial scale. The company has used its dominance in the AI chip market to increase annual revenue from $27 billion in 2022 to $216 billion last year, which the report linked to Nvidia’s growth in market value to $4.5 trillion. Nvidia’s market capitalization also briefly surpassed $5 trillion last October, before analysts and investors raised concerns that the AI boom might be overblown.
Even with a late-February quarterly report that far exceeded analyst forecasts and management’s “rosy outlook,” Nvidia’s stock price was down 6% from where it stood before those numbers came out, according to the report. After Huang’s disclosure about an anticipated doubling in backlogged chip orders, Nvidia shares edged up nearly 2% to close Monday at $183.22.
Wedbush Securities analyst Dan Ives cautioned that investors are still navigating stress rather than a steady glide path, saying it is “just a white-knuckle period for the technology industry.” Ives said, in comments reported alongside Huang’s presentation, that he expects Nvidia’s market value to eclipse $6 trillion during the next year or so.
The company’s growth outlook, however, faces competition and policy constraints, the report said. Analysts expect Nvidia’s revenue to surpass $330 billion for the upcoming year, but Nvidia is also facing its first serious challenges in the AI chip market as other companies such as Google and Meta seek to develop their own processors.
Huang described potential headwinds as part of Nvidia’s push to keep its instrumental role in AI. He said the company needs continued demand for chips that power products such as OpenAI’s ChatGPT and Google’s Gemini, and he also pointed to an expansion into inference processors for a growing set of use cases.
As Nvidia transitions more directly into inference-focused products, the report said it struck a multi-billion dollar licensing deal with Groq that included the hiring of Groq’s top engineers. “The inference inflection has arrived,” Huang said, adding that Nvidia “isn’t going to cede any market share to Google or Meta,” in Ives’s reported assessment of Nvidia’s competitive posture.