"Google Cloud" and "Intel" have announced the expansion of their multi-year cooperation in the field of artificial intelligence infrastructure.
Modern.az reports that "TechCrunch" platform has published information about this. Under the new agreement, "Google" Cloud will continue to use Intel's latest generation processors, and the parties will further deepen joint chip development.
According to the company, "Google Cloud" has been using Intel's "Xeon" processors for many years, and within the framework of this cooperation, new generation chips, including "Intel Xeon" 6, will be applied in artificial intelligence, cloud, and inference (model execution) processes. In addition, the parties will expand the joint development of special infrastructure processors (IPUs) used to manage workloads more effectively in data centers. These chips increase performance and energy efficiency by reducing the load on CPUs. This cooperation, which began in 2021, will particularly focus on the creation of special ASIC-based IPUs.
Intel's CEO Lip-Bu Tan stated that artificial intelligence infrastructure is not only composed of GPUs: "Scaling AI systems requires a balanced approach. CPUs and IPUs play a key role in this process." The recent increase in demand for CPUs in the industry also increases the importance of this cooperation. While GPUs are mainly used in training AI models, CPUs are critical for model execution and general system infrastructure. For this reason, technology companies have started to pay more attention to the CPU field.
