Intel will design CPUs with Nvidia NVLink in return for $5 billion investment

The impact for enterprises in terms of chip availability, pricing and functionality is unclear, but it all-but kills worries about Intel’s near-term viability.

Intel will collaborate with Nvidia to design CPUs with Nvidia’s NVLink high-speed chip interconnect, it said Thursday — just months after committing to co-develop a competing interconnect, UALink, with AMD, Broadcom, and other tech companies.

The two have also agreed to “jointly develop multiple generations of custom data center and PC products,” they said in a joint statement.

Nvidia will invest $5 billion in Intel stock as part of the agreement, they said.

[ Related: More Intel news and insights ]

The focus on Nvidia’s NVLink raises questions about what will happen to the multi-vendor commitment announced in April for UALink. That effort to deliver an alternative to NVLink was supported by 75 companies, including Intel, AMD, Broadcom, Cisco, Google, HPE, Meta, Microsoft and Synopsys.

Anshel Sag, a principal analyst with Moor Insights & Strategy, said it’s unclear whether Intel will continue to support UALink in the long term. Intel could continue to push the technology in the short term “because they still have to ship product until the partnership with Nvidia materializes.”

Gaurav Gupta, VP Analyst at Gartner, said both interconnect technologies will have their place in the market: “Even players that are part of UALink consortium, like Broadcom, don’t exclusively just use UALink,” he said.

[ Related: More Nvidia news and insights ]

Initial reaction to the chipmakers’ announcement was mostly favorable. Intel shares initially surged on the news but then retreated.

Sag said the Nvidia Intel deal “is simply so unprecedented that it’s hard to process to be honest. But I think it’s a clear indication that Nvidia wants Intel to be around and doesn’t see it as a near term or long-term threat in AI. Strategically, having Intel around keeps AMD on its toes too.”

The deal offers pros and cons for IT buyers across the board.

“On the client side, I really don’t know what this means for Intel’s GPU division, but I imagine that maybe it makes Intel + Nvidia solutions better, of which there are already very many. Arc‘s future seems even cloudier than before. Even though I don’t think Intel will abandon GPU IP, I do wonder whether that investment continues for discrete [graphics chips] at this point,” Sag said.

Manufacturing consent

While the companies said that Intel will manufacture the co-developed chips, they said nothing about Nvidia shifting some of its other chip lines from TSMC to Intel fabs.

“I’m also surprised that there’s no foundry component here at all, which I would imagine could help Nvidia drive down costs long term and help prop up Intel.”

Nevertheless, Info-Tech Research Group advisory fellow Scott Bickley sees opening up a second source of supply as one of Nvidia’s motivations for the deal, albeit one that will take some time to come to fruition: “Currently, its sole supplier dependency is on TSMC, which has just materially raised its prices,” he said.

Although companies bitterly fighting and then aligning is a repeating theme in Silicon Valley’s history, Sag said he found the new Nvidia-Intel alliance noteworthy. “Remember that ten years ago, they were suing each other,” he said, referring to a case settled with an out-of-court licensing deal over a decade ago.

In their joint statement today, both CEOs offered the expected self-congratulations.

Nvidia CEO Jensen Huang said, “This historic collaboration tightly couples Nvidia’s AI and accelerated computing stack with Intel’s CPUs and the vast x86 ecosystem — a fusion of two world-class platforms. Together, we will expand our ecosystems and lay the foundation for the next era of computing.”

Intel CEO Lip-Bu Tan said, “Intel’s leading data center and client computing platforms, combined with our process technology, manufacturing and advanced packaging capabilities, will complement Nvidia’s AI and accelerated computing leadership to enable new breakthroughs for the industry.”

As recently as last year, Intel appeared to be struggling and its long-term viability was unclear. But Intel’s fortunes appear to have sharply improved recently. Last month saw two significant investments to help Intel: a $2 billion boost from Softbank and an $8.9 billion investment from the U.S. government.

Moor’s Sag said that he does not expect much to change any time soon.

“This is not an immediate product roadmap change. This is going to be years out,” Sag said. He added that he doubts this alliance negates the GPU availability problems in the IT generally, as well as Oracle’s challenges with getting sufficient chips.

Sag also doubted that the Nvidia-Intel move would impact pricing. “The pricing impact is likely negligible because you still are buying an Nvidia GPU,” Sag said.