Jakarta, INTI - Google and Intel announced a partnership to strengthen artificial intelligence (AI) infrastructure, specifically through Google Cloud services. Reported by TechCrunch, through the agreement, Google Cloud will continue to utilize Intel Xeon processors, including the latest generation Xeon 6, to support AI workloads, cloud computing, and inference processes.
In addition, the two companies will also expand their joint development of dedicated infrastructure processing units (IPU), which accelerate and manage tasks in data centers by shifting workloads away from the CPU.
ASIC-Based IPUs Development
This chip development partnership, which began in 2021, will focus on application-specific integrated circuit (ASIC)-based IPUs designed specifically for AI infrastructure needs. Intel did not disclose the investment value of the partnership.
This partnership comes amid growing industry demand for processors, particularly CPUs. While GPUs are widely used to train AI models, CPUs remain a critical component in running models and supporting the overall AI infrastructure.
Intel Chief Executive Officer, Lip-Bu Tan, believes that AI is changing the way infrastructure is built and developed. He believes that AI development requires a balanced system, with the CPU and IPU playing a crucial role in delivering performance, efficiency, and flexibility.
In recent months, several technology companies have begun to refocus their attention on CPU development amid global chip supply constraints. One of them is Arm Holdings, which recently introduced the Arm AGI CPU, the company's first independently developed chip.
Conclusion
Google and Intel have partnered to strengthen AI infrastructure through Google Cloud by leveraging Xeon processors, including the latest generation Xeon 6, and developing ASIC-based infrastructure processing units (IPUs). This partnership emphasizes the critical role of CPUs in running and supporting AI systems, even though GPUs are dominant for model training. Intel CEO Lip-Bu Tan believes AI is driving the development of a more balanced infrastructure between CPUs and IPUs, amidst the growing need for chips and global supply constraints, which has also prompted other companies like Arm Holdings to develop their own processors.
Read more: OpenAI has Postponed the Stargate Data Center Project in the United Kingdom