Nvidia CEO Jensen Huang said Monday that the artificial intelligence boom is moving into a new phase, which he described as an “inference inflection.” Speaking at an event in San Jose, California, Huang predicted Nvidia will be dealing with about $1 trillion in backlogged orders for its chips by the end of the year.

Huang’s remarks were delivered during a talk that lasted more than two hours onstage, as he laid out how he says Nvidia’s processors became central components for AI deployments. He also reiterated his broader message that the AI buildup remains in its infancy.

“We reinvented computing, just like the PC (personal computer) revolution and the internet revolution,” Huang said, adding that “We are now at the beginning of a new platform change.” He tied the next wave of demand to inference, which he described as the step in which AI systems take what they have learned and produce responses.

In the framework Huang used at the event, inference chips help generate outputs such as written text or images more efficiently than the processors used while large language models were being built. He said, “The inference inflection has arrived.”

Nvidia targets $1T in chip orders, despite cooler stock trading

Huang said Nvidia’s anticipated $1 trillion backlog would amount to a doubling of his estimate from a year ago. The prediction came as Nvidia’s once-rapid share gains have cooled since the company briefly became the first to surpass a $5 trillion market value last October amid worries that the AI momentum might be overstated.

In commentary carried by the Associated Press, Wedbush Securities analyst Dan Ives said the technology industry is in a “white-knuckle period.” The report said Nvidia’s stock was down 6% from where it stood before the company released a quarterly report in late February that exceeded analyst forecasts, even as management provided a “rosy outlook.”

After Huang’s disclosures about backlogged chip orders, the report said Nvidia shares edged up nearly 2% to close Monday at $183.22.

Competition, trade barriers shape Nvidia’s next push

Beyond the near-term order outlook, the Associated Press report said analysts expect Nvidia’s revenue to surpass $330 billion for the upcoming year. It also said Nvidia’s first serious challenges in the AI chip market are emerging as large technology companies including Google and Meta Platforms try to develop their own processors.

The report said Nvidia’s growth has also been held back by U.S. security and trade barriers that have impeded the company’s ability to sell advanced chips in China.

Huang said he envisions Nvidia maintaining its role in AI by feeding the demand for chips used by chatbots such as OpenAI’s ChatGPT and Google’s Gemini. He also said Nvidia wants to expand its reach into the market for inference processors.

Nvidia partnership supports shift into inference

To help navigate that transition, the Associated Press reported that Nvidia struck a multi-billion dollar licensing deal with market specialist Groq. The report said the agreement included hiring Groq’s top engineers.

Ives, in the Associated Press report, said Nvidia is “isn’t going to cede any market share to Google or Meta.” He said he believes Nvidia’s market value will exceed $6 trillion during the next year or so.