Intel and Google officially announced an expanded long-term collaboration to advance next-generation AI infrastructure.

Intel announced on the 10th that it is pursuing a strategic partnership with Google to further strengthen the role of CPUs and application IPUs (Infrastructure Processing Units) to efficiently operate increasingly complex cloud environments amidst the rapid spread of AI adoption.

As AI technology is utilized on a large scale, data processing structures are becoming increasingly heterogeneous.

Accordingly, the importance of the CPU is being highlighted again to coordinate the entire system and maintain stable performance.

In line with this trend, Intel and Google are continuously optimizing infrastructure based on Intel Xeon processors across multiple generations and strengthening their collaboration with the goal of improving performance, power efficiency, and Total Cost of Ownership (TCO).

Google Cloud is already operating instances optimized for various workloads, including C4 and N4 instances equipped with the latest Intel Xeon 6 processors.

These instances support a wide range of tasks, including large-scale AI training orchestration, latency-sensitive inference, and general-purpose computing, continuously expanding the scope of Intel CPU utilization.

Both companies are also expanding the joint development of ASIC-based IPUs.

IPU is Networking, Storage & MiddoIt is an accelerator that separates and processes infrastructure functions previously handled by the CPU, such as security, and is considered a key element in increasing performance predictability and efficiency in hyperscale environments.

This allows cloud providers to enhance scalability without increasing system complexity.

Intel Xeon CPUs and IPUs are integrated in a complementary architecture to form a platform that provides a balance of general-purpose computing and specialized infrastructure acceleration capabilities.

This structure plays a crucial role in meeting the performance and flexibility required at various stages, such as AI training, inference, and deployment.

Intel CEO Rip-Bhutan stated, “The spread of AI is changing the very way infrastructure is designed,” emphasizing the importance of a balanced system configuration rather than a simple accelerator-centric one.

He explained that CPUs and IPUs are key pillars meeting the demands of the latest AI workloads.

Amin Bahdat, Senior Vice President of Google’s AI Infrastructure, also stated, “CPU and infrastructure acceleration technologies are essential elements throughout the entire process of AI systems,” and assessed that the partnership with Intel, which has lasted for nearly 20 years, has served as the foundation for meeting Google’s performance and efficiency requirements.