- Nvidia will invest $5 billion in Intel (about a 4% stake) and co-develop tightly integrated CPU-GPU products for PCs and AI data centers.
- The plan pairs Intel x86 CPUs with Nvidia RTX GPU chiplets in new SoCs and uses custom Intel CPUs as hosts for Nvidia rack-scale systems via NVLink.
- The tie-up fuels speculation Intel could scale back or exit its Arc discrete GPU effort, though Intel says its Arc and integrated GPU roadmaps remain intact.
- The partnership could squeeze AMD while leaving open whether Nvidia will ever become a meaningful customer of Intel Foundry.
Read More
In September 2025, Intel and Nvidia entered into a $5 billion equity partnership. Nvidia will acquire Intel common stock (≈4% ownership at $23.28/share) and work with Intel on co-developing PC and AI data center products that tightly combine Intel’s x86 CPU architecture with Nvidia’s RTX GPU chiplets and NVLink interconnect technology.
Central to the collaboration are two product lines: a) PC/consumer devices built around system-on-a-chip (SoC) designs integrating Intel CPUs and Nvidia RTX GPU chiplets—targeting laptops and mainstream PCs; and b) custom Intel CPUs for Nvidia rack-scale AI platforms, which will replace or supplement existing Arm-based hosts like Grace in certain high-performance server offerings.
Given Intel’s history of exiting non-core businesses under new CEO Lip-Bu Tan—such as memory, Ethernet switches, Bitcoin chips, and even partially spinning off Altera and selling off Mobileye—the Nvidia collaboration has triggered speculation that Arc/Xe discrete graphics, or at least its high-end component, may be among the units to be curtailed. The Motley Fool explicitly argues the graphics business is the most likely candidate to be “axed.”
However, Intel has pushed back. The company has stated that it will continue its existing roadmap for Arc, integrated graphics, and discrete GPUs, calling the partnership “additive” rather than substitutive. Intel insists that the deal does not replace its own GPU technology or negate investment in discrete graphics.
Strategic implications are wide and include:
- AMD is under duress. The combined Intel-Nvidia product architecture could erode AMD’s competitive advantage in both APUs and discrete graphics.
- The foundry business remains uncertain: while Nvidia has evaluated Intel Foundry, the current agreement is design-focused, with no firm commitment for Nvidia to manufacture via Intel’s fabs. Losses in Intel Foundry (approximately $13 billion in FY 2024) remain a burden.
- Intel may face dependency risk. Relying on Nvidia for graphics innovation could save cost but increase dependency, possibly limiting Intel’s margins and strategic autonomy.
Open questions going forward include:
- Will Intel eventually discontinue or significantly downscale its high-end discrete graphics (Arc/Xe) in favor of the new Intel-Nvidia SoCs?
- How will Intel balance its GPU roadmap, including integrated graphics (iGPU), discrete GPUs (dGPU), and this new hybrid RTX-chiplet path, in terms of investment, engineering effort, and product segments?
- Will Nvidia eventually use Intel’s foundry/fab facilities to manufacture GPU die or packaging, deepening the partnership beyond just chip design and packaging?
- How will OEMs, customers, regulatory bodies, or antitrust concerns respond to this tightening of cooperation between competitors?
Supporting Notes
- Nvidia is investing $5 billion in Intel common stock, giving Nvidia about a 4 % stake.
- The joint products will include Intel x86 CPUs integrated with Nvidia RTX GPU chiplets in PC laptops via system-on-chip designs; and custom Intel CPUs used in Nvidia rack-scale AI platforms using NVLink.
- Intel’s discrete graphics brand Arc (Xe-based GPUs) is a direct competitor to Nvidia and AMD; the upcoming partnership may reduce demand for Arc’s high-end offerings.
- Intel has publicly committed to continuing its GPU products and roadmap, including discrete and integrated graphics, despite the partnership.
- The Intel Foundry division has experienced steep losses (~US$13 billion in 2024), and lacks major external customers—putting it under pressure.
- AMD’s 2025 report identifies the Intel-Nvidia partnership as a material risk to AMD’s business: expected intensification of competition in APUs and pricing pressure.