Google has committed to using multiple generations of Intel Xeon 6 processors in its AI data centers for training and inference workloads, expanding a partnership that dates back nearly three decades to Google’s earliest server infrastructure. Intel shares gained nearly 5% on the announcement, CNBC reported.

The Deal

No financial terms were disclosed, and the companies did not provide a timeline. Google’s chief technologist for AI infrastructure, Amin Vahdat, said in a statement that Intel’s “Xeon roadmap gives us confidence that we can continue to meet the growing performance and efficiency demands of our workloads,” according to CNBC.

Intel makes the latest Xeon 6 on its most advanced 18A manufacturing technology at its Arizona chip fabrication plant, which opened last year. The deal gives Intel a stronger position in AI data centers, a market dominated by Nvidia’s GPUs.

CPUs as the Next Bottleneck

The announcement arrives as the industry reassesses CPU importance in AI infrastructure. Nvidia’s head of AI infrastructure, Dion Harris, told CNBC in March that CPUs are “becoming the bottleneck” as agentic workloads push compute requirements beyond GPUs alone.

“Scaling AI requires more than accelerators,” Intel CEO Lip-Bu Tan said in a statement about the deal, per CNBC. The framing positions Xeon not as a GPU competitor but as essential infrastructure alongside accelerators.

IPU Collaboration

Google and Intel also reiterated their ongoing collaboration on Infrastructure Processing Units (IPUs), a partnership that started in 2022. The IPU is a programmable accelerator used to offload networking, storage, and security functions from host CPUs. Google told CNBC that the IPU was “a first-of-its-kind chip” when the companies first collaborated on it four years ago.

Intel’s Broader Momentum

Intel has nearly tripled in share price over the past year, fueled by a 10% government equity stake sold in August 2025 and a subsequent $5 billion Nvidia stake. This week, CEO Lip-Bu Tan posted on LinkedIn that Elon Musk has tapped Intel to design, fabricate, and package custom chips for SpaceX, xAI, and Tesla at the Terafab project in Texas, though no financial details were provided, per CNBC.

The Infrastructure Layer Shift

The Google deal reinforces a shift in how AI infrastructure is evaluated. The GPU shortage narrative that dominated 2024 and 2025 is giving way to a systems-level view where CPUs, IPUs, networking, and power delivery are all bottlenecks. For Intel, which spent years losing ground in consumer chips, the pivot to AI infrastructure components alongside Google validates a strategy that Lip-Bu Tan has staked his turnaround on. The open question is whether Intel’s 18A manufacturing can deliver at the volume and yield rates that Google’s scale demands.