Nvidia Invests $5 Billion in Intel to Co-Develop AI Chips for PCs and Data Centers

Nvidia and Intel have announced a groundbreaking partnership that marks a dramatic shift in their long-standing rivalry, with both companies committing to jointly develop multiple generations of custom x86-based products. The collaboration includes the creation of x86 RTX System-on-Chips (SoCs) for consumer gaming PCs, where Intel will design and manufacture CPUs tightly integrated with Nvidia’s RTX GPU chiplets using the high-bandwidth NVLink interface—offering up to 14 times more bandwidth and lower latency than traditional PCIe connections. These SoCs will target thin-and-light gaming laptops and small form-factor PCs, competing directly with AMD’s APUs. Unlike Intel’s earlier Kaby Lake-G effort, which used PCIe and separate GPU memory, the new chips will feature uniform memory access (UMA), allowing both CPU and GPU to share a single memory pool. In addition, Intel will produce custom x86 data center CPUs for Nvidia, designed specifically for integration into Nvidia’s AI infrastructure platforms. These CPUs will also leverage NVLink Fusion technology, enabling faster, more efficient communication between CPUs and GPUs than PCIe-based systems. This move strengthens Nvidia’s position in hyperscale and enterprise AI markets, where performance and scalability are critical. As part of the deal, Nvidia will invest $5 billion in Intel common stock at $23.28 per share, acquiring approximately a 5% ownership stake. The investment, subject to regulatory approval, follows a $9 billion U.S. government stake and a $2 billion investment from SoftBank, all aimed at stabilizing Intel amid financial strain and intense competition from AMD and TSMC. Intel’s stock surged 33% in premarket trading, reflecting strong market confidence in the partnership. The collaboration does not signal a retreat from Nvidia’s other roadmaps. The company reaffirmed its commitment to its Arm-based Grace Blackwell and Vera CPUs, as well as internal projects. It remains unclear whether Intel’s foundries will produce the new chips, though Intel’s push to regain manufacturing leadership—using its own Intel 3 and upcoming 18A process nodes—suggests it may play a key role, especially for data center products. The partnership also raises questions about the future of Intel’s Arc GPU line. With Nvidia providing the GPU technology for new SoCs, Intel may increasingly focus on its CPU and manufacturing strengths rather than competing in discrete graphics. Nvidia CEO Jensen Huang hailed the alliance as a “historic collaboration” that fuses Nvidia’s AI and accelerated computing leadership with Intel’s x86 ecosystem and manufacturing prowess. Intel CEO Lip-Bu Tan praised the investment as a vote of confidence in Intel’s innovation and long-term strategy. The announcement comes amid a broader industry trend of consolidation and strategic partnerships, driven by the need to accelerate AI development and manage rising R&D and fabrication costs. While the new products are still in early stages and likely won’t reach market for at least a year, the partnership signals a new era of cross-platform integration and could reshape competition in both consumer and data center markets. A joint press conference hosted by both CEOs will provide further details. The collaboration, while transformative, remains subject to regulatory scrutiny and market dynamics. Still, it represents a pivotal moment in semiconductor history, uniting two industry giants to drive the next wave of computing innovation.