Intel, Google Expand AI Partnership on Xeon, IPUs

Intel and Google announced an expanded multiyear partnership on Thursday to advance AI infrastructure, focusing on increased deployment of Intel's Xeon processors in Google Cloud and joint development of custom infrastructure processing units.

The agreement builds on nearly two decades of collaboration between the companies. Google Cloud will deploy Intel's latest Xeon 6 processors across its global data centers, powering workload-optimized C4 and N4 instances. These instances support a range of AI tasks, including training coordination, latency-sensitive inference, and general-purpose computing. Officials from both companies stated that the Xeon 6 chips provide more than double the total cost of ownership benefits compared to previous configurations, combining performance gains with improved energy efficiency.

In addition to the processor deployment, Intel and Google will deepen their codevelopment of Infrastructure Processing Units, or IPUs. These custom, programmable accelerators are designed to handle networking, storage, and security functions, offloading them from host CPUs to enhance utilization and predictability in large-scale AI environments. The partnership emphasizes...