Intel inks multiyear data center chip partnership with Google

Google LLC will use many future iterations of Intel Corp.’s series. of Xeon as part of the partnership announced today.

Shares of the chip maker closed 4.7% higher on the news.

Google will use the chips in its cloud environment. They will power artificial intelligence models and general purpose workloads.

Google Cloud already uses Xeon 6, Intel’s new series of central processing units, to run some of its general-purpose C4 instances. Virtual machines can reach a maximum clock speed of 3.9 gigahertz when all the cores of the underlying CPU are active. That frequency can go up to 4.2 gigahertz if only the fastest cores are online.

Google’s C4 instances use a variant of Xeon 6 called Granite Rapids. It is based on a core design called P-core which includes many AI-focused optimizations. Some of that is being optimized for AMX, a set of machine language extensions in which Intel chips express math. AMX accelerates a calculation called multiply-encode AI models that run continuously during forecasting.

Intel also offers a second batch of Xeon 6 chips called Sierra Forest. Those CPUs are based on a core design called E-core that trades some of the P-core functionality for improved performance.

Intel released its most advanced E-core processor in March. It includes 288 cores, or 160 more than the largest Granite Rapids processor. The chip is based on the company’s latest Intel 18A manufacturing process, which offers up to 15% better performance per watt than the Intel 3 node supporting previous Xeon 6 chips.

“Scaling AI requires more than accelerators – it requires scalable systems,” said Intel Chief Executive Lip-Bu Tan. “CPUs and IPUs are critical to delivering the performance, efficiency and flexibility on demand for modern AI workloads.”

The companies’ new partnership also extends to Intel’s IPU, or infrastructure processing unit, product family. The chips in the array are configured to perform infrastructure management tasks such as encrypting data traffic and connecting storage hardware. IPUs offload those tasks to the server’s CPU, leaving more computing capacity for users’ workloads.

Intel and Google plan to expand their “co-development of ASIC-based IPUs.” An ASIC, or application-specific integrated circuit, is a processor designed from the ground up for a specific set of use cases. That suggests Google will use IPUs optimized for its cloud data centers.

The contract is a much-needed win for Intel, which faces increasing competition in the server CPU market. Last month, rival Arm Holdings plc first released its first ready-made processor for data centers. The 136-core AGI CPU was developed in partnership with Meta Platforms Inc., which will use the chip to power its internal AI infrastructure.

Photo: Intel