Analysis of winners and losers from the Intel-Nvidia CPU-GPU partnership including processor market implications.

Nvidia is acquiring a 5% stake in Intel for $5 billion, dipping into its $56 billion bank account to become the second-largest shareholder after the federal government's recent investment. The deal does not provide Nvidia with a board seat or operational control—it is simply a stock purchase focused on Intel's products division, not its foundry business. Most significantly, the investment provides Nvidia greater access to the x86 ecosystem, which is crucial for the enterprise data center market, while giving Intel access to GPUs in high demand and an opportunity to move its CPU products forward. Technologically, the deal brings the two companies much closer together, about as close as two companies can get short of a merger, which analysts do not believe will happen. Anshel Sag, principal analyst with Moor Insights & Strategy, stated, "No, I don't think that's the path. But it could be during this administration, if there were a time and place."

In announcing the deal, Nvidia CEO Jensen Huang emphasized that future Intel chips would have Nvidia GPUs baked into them instead of Intel's own GPU technology. Alvin Nguyen, senior analyst with Forrester Research, views this as "a safe investment and partnership to get them some PR, greater access to the x86 market, and provide their GPU technology to be included in APUs and SoCs that can reach markets they are not established in," though he acknowledges regulatory hurdles and complex geopolitics surrounding such arrangements. The server business will also be impacted by this strategic partnership.

AMD emerges as the clear loser from this alliance. Previously, AMD held a competitive advantage by offering CPU and GPU combinations that Intel and Nvidia lacked individually—an advantage visible in supercomputers like Frontier and El Capitan, which are all-AMD designs. Now that the two companies are allied, they will have a competitive offering in the near future. Additionally, the future of Intel's Jaguar Shores AI accelerator, based on its GPU technology and the Gaudi AI accelerator, has become uncertain. As Nguyen noted, "Nvidia already has solutions here, and it doesn't make sense for Intel to work on a redundant product that needs to be marketed over an established one."

A significant outcome is Intel's adoption of Nvidia's proprietary NVlink high-speed interconnect protocols. Jack Gold of J. Gold Associates wrote, "This means that Intel has essentially determined its ability to compete head-to-head with Nvidia in the current large-scale AI marketplace, despite its best efforts, have mostly failed." Nvidia already uses Xeon data center chips in its largest systems, with x86 chips providing most of the controls and pre-processing required for large-scale GPU racks. The arrangement also leaves a question mark over Nvidia's Arm CPUs, likely limited to "niche areas." Gold observed that with this announcement, it "now seems to admit that working with an established player like Intel is a better long-term bet for CPUs." When asked whether Nvidia would continue its Arm efforts, Sag responded, "Nvidia has made it abundantly clear that Arm is moving forward as planned, especially since Nvidia is still building Arm for its own CPUs."

Such a deal would have seemed impossible 15 to 20 years ago, given the intense acrimony between the two companies. Today, however, Nvidia's Huang is partners with Intel's latest CEO, Lip-Bu Tan. Jack Gold captured the essence of the shift: "Sometimes enemies become best of friends when there is money to be made."