On April 9, Intel and Google jointly announced that the two parties will launch an in-depth, multi-year cooperation around Xeon CPUs and customized infrastructure processing units (IPUs) to jointly promote the development of next-generation artificial intelligence and cloud infrastructure.

Core content of cooperation

This cooperation covers two core areas:

1. Google Cloud will continue to use multiple generations of Intel Xeon processors, including the latest Xeon 6 chips, for AI training, inference and general computing workloads

2. Intel CPUs are officially incorporated into Google’s AI workload system, marking a re-examination of the strategic value of CPUs in AI infrastructure

Customized IPU joint development

Both parties expand joint development of custom ASIC-based IPUs, starting in 2021, to offload network, storage and security functions from the main CPU

Improve CPU utilization, improve efficiency, and achieve more predictable performance in hyperscale AI environments

This agreement is a substantial deepening of a long-standing relationship between Google and Intel that dates back to when Google first built server racks nearly three decades ago. Strategic context: CPUs return to spotlight in AI race

The collaboration comes as CPUs are returning to the spotlight in the next phase of the AI ​​race. Dion Harris, head of Nvidia's AI infrastructure, told the media in March this year that the CPU is "becoming the bottleneck" as agent workloads push computing power requirements beyond the graphics processor.

Intel CEO Lip-Bu Tan emphasized in the statement: "Artificial intelligence is reshaping how infrastructure is built and scaled. Scaling artificial intelligence requires more than just incubators, but balanced systems. CPUs and IPUs are critical to delivering the performance, efficiency, and flexibility required for modern artificial intelligence workloads."

Amin Vahdat, senior vice president and chief technology officer of Google AI Infrastructure, said: "CPU and infrastructure acceleration remain the cornerstone of artificial intelligence systems - from training orchestration to inference and deployment. Intel has been a trusted partner for nearly two decades, and their Xeon processor roadmap gives us confidence that we can continue to meet the growing performance and efficiency needs of our workloads." Market reaction and industry impact

For Intel, this cooperation is an important opportunity to enter the AI ​​chip market dominated by Nvidia. Intel has made intensive moves recently, including the U.S. government’s purchase of 10% of Intel’s shares, Nvidia’s $5 billion investment in Intel, and Elon Musk’s commissioning of Intel to design and manufacture custom chips for SpaceX, xAI, and Tesla. Neither party disclosed the financial terms of the cooperation or the timetable for the agreement.