Google and Intel Forge Multiyear CPU and Custom IPU Partnership for AI Infrastructure

Google and Intel Forge Multiyear CPU and Custom IPU Partnership for AI Infrastructure

April 9, 2026

In a significant move to shape the future of artificial intelligence infrastructure, Intel and Google have announced a comprehensive, multiyear agreement that solidifies their long-standing partnership. The deal underscores the critical and evolving role of traditional processors in the era of generative AI, where balanced system architecture is becoming paramount for scaling complex workloads.

The collaboration centers on two key pillars: procurement and co-development. Google has committed to deploying multiple future generations of Intel's Xeon 6 CPUs across its Google Cloud Platform (GCP). Specifically, Intel's Xeon processors will power C4 and N4 instances on GCP, supporting a broad spectrum of workloads from large-scale AI training and inference to general-purpose computing. This commitment builds upon Intel's launch of its Xeon 6 portfolio, which the company has promoted as delivering an average performance improvement of 1.4 times compared to its previous generation across enterprise workloads.

Beyond standard CPUs, the partners will jointly engineer custom, ASIC-based Infrastructure Processing Units (IPUs), extending work that began with the 'Mount Evans' project announced in 2021. These specialized chips are designed to offload infrastructure management tasks from main server CPUs, freeing up valuable processing cycles for customer applications. The financial terms and specific timeline for this expansive agreement were not disclosed.

Executive commentary highlighted the strategic necessity of the partnership. "AI is reshaping how infrastructure is built and scaled," stated Intel CEO Lip-Bu Tan. "Scaling AI requires more than accelerators - it requires balanced systems. CPUs and IPUs are central to delivering the performance, efficiency, and flexibility modern AI workloads demand."

Echoing this sentiment, Amin Vahdat, SVP and chief technologist for AI infrastructure at Google, noted, "CPUs and infrastructure acceleration remain a cornerstone of AI systems—from training orchestration to inference and deployment. Intel has been a trusted partner for nearly two decades, and their Xeon roadmap gives us confidence that we can continue to meet the growing performance and efficiency demands of our workloads."

The deal arrives amid a notable resurgence in demand for data center CPUs, driven by the computational needs of agentic AI. Intel itself recently acknowledged a market miscalculation, with CFO David Zinsner admitting the company underestimated this demand surge. Intel is now reallocating manufacturing capacity to address a shortage it expects to persist into its 2026 fiscal year, anticipating the tightest supply in the first quarter before conditions gradually improve. This partnership with Google, a leading hyperscaler, provides Intel with a stable, high-volume deployment channel for its core data center products during this critical period.

Source: datacenterdynamics