Google expands partnership with Intel for AI chips

Technology giant Google has expanded its strategic partnership with Intel by committing to use multiple generations of Intel's central processing units (CPUs) to power its artificial intelligence data centers, as announced on Thursday, November 17, 2025. This move aims to diversify Google's AI hardware supply chain and bolster Intel's position in the competitive AI chip market, which has been dominated by Nvidia. The announcement, made from Intel's advanced packaging facility in Chandler, Arizona, signals a deepening of a nearly three-decade-long collaboration between the two companies, dating back to Google's earliest server infrastructure. The core of the expanded deal involves Intel's newest Xeon 6 CPUs, which will now be deployed to run both AI training and inference workloads within Google's vast data center network. Amin Vahdat, Google's chief technologist for AI infrastructure, stated that Intel's Xeon roadmap provides the confidence needed to meet the growing performance and efficiency demands of Google's AI operations. While no financial terms or specific timelines were disclosed, the partnership underscores a significant industry shift: the central processing unit is regaining prominence as a critical component in scaling advanced AI systems. Industry leaders, including Nvidia's head of AI infrastructure, have noted that CPUs are becoming a potential bottleneck as AI workloads evolve, necessitating more balanced computing architectures. This partnership arrives at a pivotal moment for Intel, which has been striving to regain its footing in the high-stakes semiconductor industry. Recent investments, including a 10% stake sale to the U.S. government in August and a subsequent $5 billion stake purchase by Nvidia, have fueled a near-tripling of Intel's stock over the past year. The deal with Google also highlights the complex, multi-faceted nature of modern chip alliances. Alongside the CPU agreement, Google and Intel reiterated their ongoing collaboration on Infrastructure Processing Units (IPUs), specialized chips designed to offload networking and storage tasks from main CPUs. Notably, Google continues to develop its own custom silicon, such as the Tensor Processing Unit (TPU) for AI and the Arm-based Axion CPU, indicating a hybrid strategy of partnership and in-house development to secure its technological future.