Intel and Google launch a new phase of collaboration in AI

Intel and Google have announced an expansion of their long-standing collaboration in the development of artificial intelligence infrastructure and cloud services. The companies intend to jointly create next-generation solutions, focusing on a combination of central processing units and specialized accelerators.

As part of the partnership, Intel Xeon processors will continue to be used in Google Cloud's global infrastructure for AI training and inference tasks, as well as for general-purpose computing. Special attention is being paid to improving performance, energy efficiency, and reducing the total cost of ownership of the infrastructure.

At the same time, the companies are strengthening their joint development of specialized ASIC solutions — IPU infrastructure processors. These chips take on network communication, data storage, and security tasks, relieving the central processing units and increasing the overall efficiency of data centers.

Intel emphasizes that modern AI systems require not only accelerators but also a balanced architecture where CPUs play a key role. In turn, Google notes that further AI development is impossible without optimizing the entire infrastructure — from model training to deployment.

The expansion of cooperation is aimed at creating more flexible and scalable AI solutions that will enable companies and developers to handle growing workloads more effectively and accelerate the adoption of new technologies.

Source: intc.com