Google and Intel deepen AI infrastructure partnership

In a significant move, Google and Intel revealed an extended partnership on Thursday, aiming to bolster their collaboration in AI infrastructure over the coming years. This partnership will see Google Cloud continue to leverage Intel's AI capabilities, particularly through the use of Intel’s Xeon processors, including the cutting-edge Xeon 6 chips, for various AI, cloud, and inference operations. For decades, Google Cloud has relied on Intel's Xeon processors, and this latest agreement underscores a commitment to further innovation. A key aspect of the partnership involves the co-development of specialized infrastructure processing units (IPUs). These IPUs are designed to enhance data center efficiency by offloading specific tasks from traditional CPUs, thereby streamlining operations. Initiated in 2021, this collaboration will now focus on developing custom application-specific integrated circuits (ASICs) for IPUs. However, Intel has not disclosed any pricing details related to this expanded partnership. As the demand for CPUs surges, the industry is experiencing a notable shortage. While GPUs have been the go-to for training AI models, CPUs play a vital role in executing these models effectively and supporting the broader AI infrastructure. Intel's CEO, Lip-Bu Tan, emphasized the transformative impact AI is having on infrastructure development. He stated, "AI is reshaping how infrastructure is built and scaled. 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." With the growing chip shortage, more companies are shifting their focus to CPUs. Recently, Arm Holdings, a subsidiary of SoftBank, introduced the Arm AGI CPU, marking its first self-produced chip amid global supply constraints.