Jensen Huang explains why Nvidia's latest partnership with OpenAI is different

According to CNBC’s Jim Cramer, Jensen Huang, the CEO of Nvidia, said the company has entered its first ever direct partnership with OpenAI, changing how the ChatGPT maker will get its hardware.

Jensen explained the deal on Tuesday during the CNBC Investing Club’s Monthly Meeting at the New York Stock Exchange, adding that OpenAI will now buy Nvidia systems straight from the company instead of through cloud providers.

“This is a partnership that, for the first time, OpenAI is going to buy directly from us,” Jensen said. “Usually…a cloud service provider buys from us, and they rent from a cloud service provider. And so now it’s going to be a direct partnership.”

The deal is backed by numbers that show just how large it is.Nvidia announced in September that it plans to invest up to $100 billion into OpenAI to build out artificial intelligence data centers.Both companies said the systems will require 10 gigawatts of power.

Jensen told CNBC that this amount of energy is the same as running between 4 million and 5 million GPUs. That puts OpenAI on track to run its own computing infrastructure rather than relying on someone else’s servers.

Nvidia signs direct AI infrastructure deal

The partnership comes as both Nvidia and OpenAI are driving the current AI boom. Demand for Nvidia’s chips began three years ago when OpenAI first released ChatGPT, which introduced generative AI to millions of users.

Since then, Nvidia’s market cap has more than tripled, reaching over $4 trillion and making it the most valuable company on the market. Jensen told Cramer that the OpenAI deal is “incremental” compared to Nvidia’s earlier partnerships with Oracle and CoreWeave, because this time OpenAI will run its own data centers.

“We’re going to help them build an AI infrastructure that they operate themselves … and really set them up for, you know, five years out, when they’re going to operate their own cloud anyhow,” Jensen said.

During the conversation, Jensen and Cramer also talked about China, global trade, and competition with rivals like AMD and Intel. The focus stayed on the scale of the OpenAI agreement, which changes how these massive AI operations are set up.

Intel, facing its own challenges, is preparing to show technical details about its next laptop chip called Panther Lake. Four people briefed on the matter allegedly told Reuters that the company will release the information on Thursday. Panther Lake is the first Intel chip built fully with its 18A manufacturing process, a technology the company has spent billions on. It will power high-end laptops and is part of Intel’s push to recover ground lost to AMD.

Intel prepares Panther Lake rollout

Intel ran hours of technical briefings and factory tours in Arizona last week to present the new Panther Lake microarchitecture to analysts and journalists. The briefings included details about redesigned graphics cores, processor cores, and a new AI engine.

Panther Lake will be the first large-scale Intel product made with 18A. The previous generation, called Lunar Lake, was made mostly by Taiwan Semiconductor Manufacturing Co. Intel said Panther Lake will be available in early 2026, and it will use 30% less energy while delivering a 50% boost in certain data-processing tasks.

The pressure on Intel is heavy. The company reported a $2.9 billion loss in the second quarter. It also said it could suspend its planned 14A process if it failed to secure a customer. In August, President Donald Trump called for Intel CEO Lip-Bu Tan to resign.

That sparked new investment into the company. Intel secured money from SoftBank Group and Nvidia, while the White House agreed to convert a CHIPS Act grant into a 9.9% equity stake in Intel after Tan met with Trump and officials.

The smartest crypto minds already read our newsletter. Want in? Join them.