Intel-Nvidia Deal Will Create ‘New Class Of Integrated Graphics Laptops,’ Huang Says

The plan for Intel and Nvidia to co-design new system-on-chips for laptops combined with their plan to integrate Intel CPUs into Nvidia’s rack-scale platforms for AI data centers will represent an annual market opportunity worth up to $50 billion, says Nvidia’s CEO.

Nvidia CEO Jensen Huang said Thursday that the company’s new deal with Intel will allow the two firms to create a “new class of integrated graphics laptops,” representing what he called an “underserved” market that is “largely unaddressed by Nvidia today.”

This combined with their plan to integrate Intel CPUs into Nvidia’s rack-scale AI platforms for data centers will represent an annual market opportunity worth up to $50 billion, Huang (pictured right) said in an afternoon webcast alongside Intel CEO Lip-Bu Tan (pictured left), hours after the two companies announced their “historic” partnership.

“In both cases, it expands the market for Intel very significantly and expands the market for Nvidia as well,” Huang said.

The two companies announced this morning that Nvidia plans to invest $5 billion in Intel common stock and jointly develop “multiple generations” of PC and data center products with the chipmaker. Intel’s stock price was up by more than 22 percent while Nvidia’s shares were up by more than 3.5 percent by Thursday mid-afternoon Eastern Time.

Huang said Nvidia decided to buy a $5 billion equity stake in Intel “because we thought it was going to be such an incredible investment.” He added elsewhere in the call that the Trump administration “had no involvement in this partnership at all.”

The Trump administration has played an outsized role for both companies. President Trump announced last month that the U.S. government reached an agreement to gain a 9.9 percent equity stake in Intel. Nvidia, on the other hand, has been working with the Trump administration as it seeks to gain new licenses to sell GPUs to customers in China.

Nvidia, Intel To Enable ‘New Class Of Integrated Graphics Laptops’

The joint development efforts will center around using Nvidia’s NVLink interconnect technology to “seamlessly” connect Intel’s and Nvidia’s respective chip architectures.

In the webcast, Huang and Tan provided more details about the partnership and the products Intel and Nvidia will create together, but they did not commit to a timeline for when such offerings will become available to customers.

For the PC market, Intel plans to create a “giant” system-on-chip that fuses a custom Intel x86-based CPU and Nvidia RTX GPU chiplet using NVLink to create a “new class of integrated graphics laptops that the world’s never seen before,” according to Huang.

Tan said the joint product will also feature unified memory.

Huang said this will represent a new market for Nvidia because the company has largely focused on providing discrete GPU for gaming laptops and mobile workstations in PC world. Laptops outside of those categories, on the other hand, mostly use integrated graphics, a “segment that has been largely unaddressed by Nvidia today,” he added.

“We’re going to be quite a large supplier of GPU chiplets into Intel x86 SoCs,” Huang said.

Tan did not say what impact this will have on Intel’s Arc GPU technology, which the company has used to power integrated graphics for chips like the Core Ultra 200 series. The chipmaker has been a long-term provider of integrated graphics for laptops and desktops.

Nvidia To Enable x86-Based Rack-Scale AI Platforms

For data centers, Intel plans to design a custom x86-based CPU that can integrate directly into Nvidia’s rack-scale platforms like the NVL72, which enables high-speed communication between 72 GPUs and the host CPUs using NVLink to allow them to act if they’re “one giant computer,” according to Huang.

“In the future, we will buy x86 CPUs from Intel, and we would fuse it with NVLink into our rack-scale system, so we’re going to become a very large customer of Intel CPUs,” he said.

Nvidia previously used NVLink exclusively to enable super-fast connections between its GPUs and its Arm-based CPUs in products like the GB200 NVL72 rack-scale platform. The company developed the technology in large part to provide faster chip-to-chip communication and enable other capabilities in contrast to what is possible with the PCIe standard.

But that changed in May when Nvidia announced that it is opening up NVLink for other companies to connect their CPUs with Nvidia GPUs (or conversely, their accelerator chips with Nvidia CPUs) using the interconnect technology.

At the time, Nvidia said it was working with Fujitsu and Qualcomm, whose Arm-based CPUs can be integrated with its GPUs using NVLink for rack-scale platforms.

While Huang said Nvidia remains “fully committed” to its road map for Arm-based products and added that the company has “lots of customers for Arm,” the new rack-scale platforms it will build using Intel CPUs will be aimed enterprise customers who are still running on the x86 instruction set architecture and not Arm.

“The vast majority of the world’s enterprises [are] still x86-based. They now have state-of-the-art AI infrastructure,” he said.

Tan did not say whether this design partnership will impact Intel’s push to introduce a new AI accelerator chip meant for rack-scale platforms.

“Nvidia is [the] clear leader in AI accelerated computing. Intel is the leader in the data center and client PC CPU. This collaboration brings all together for the best for the industry going forward,” he said on the webcast.

Huang, Tan Address Whether The Deal Will Boost Intel Foundry

In response to questions about whether the deal will result in Nvidia becoming a customer of the Intel Foundry contract chip manufacturing business, Huang said the partnership is “100 percent focused” on the products it’s designing with Intel.

“We’ve always evaluated Intel’s foundry technology, and we’re going to continue to do that,” he said. “But today this announcement is squarely focused on these custom CPUs.”

However, Huang said that they plan to use Intel’s Foveros 3-D packaging technology to bring together the CPU and GPU inside the system-on-chip they’re designing for laptops.

“It’s really a fabulous way of mixing and matching technology, and it’s one of the reasons why we’re going to be able to innovate so quickly and build these incredibly complex systems and deliver [them] as multi-chiplet systems packages,” he said.

The comments were made as Intel pushes to improve its foundry capabilities in a move to win over a significant customer that Tan has stated is necessary for the company to continue investing in leading-edge manufacturing capabilities in the future.

Tan said he and Huang will decide whether these jointly developed products are right for Intel Foundry as opposed to Taiwanese foundry giant TSMC, which Nvidia uses to manufacture many of its products and Intel uses to fabricate some of its products.

“We’re going to continue driving our success on the process side and then win the customer [confidence] and trust […] one step at a time,” Tan said.