Google and Intel Expand AI Chip Partnership to Power Cloud
Google and Intel Expand AI Chip Partnership to Power Cloud
Google Cloud deepens its AI infrastructure pact with Intel, expanding CPUs and IPUs to power AI workloads amid a global chip shortage.
Google and Intel are expanding a multiyear partnership focused on AI infrastructure for Google Cloud, with plans to continue using Intel Xeon processors and to jointly develop custom IPUs to accelerate data center tasks. The collaboration, which began in 2021, now places a stronger emphasis on ASIC-based IPUs designed to offload workloads from CPUs and improve efficiency for AI inference and cloud tasks. This expansion comes as the industry faces a tightening supply of CPUs, prompting firms to pursue balanced systems that can handle diverse AI workloads.
Intel is pushing a vision where CPUs and IPUs work in tandem to deliver the performance, efficiency, and flexibility modern AI workloads demand. The company’s leadership has emphasized that while GPUs remain essential for model development and training, CPUs play a critical role in running AI models in production and in general AI infrastructure. By combining Intel’s Xeon ecosystem with new IPU developments, Google Cloud aims to offer customers more scalable and streamlined AI services.
As enterprises look to scale AI across data centers, the partnership is positioned to help Google Cloud deliver consistent performance for AI workloads, from data ingestion and preprocessing to real-time inference. The collaboration will also explore further generations of custom hardware components tailored to Google’s cloud architecture, with the goal of boosting efficiency and reducing latency for AI-driven operations. Amid ongoing concerns about chip availability, the alliance underscores a broader industry trend toward diversified, purpose-built hardware alongside traditional CPUs.
