Intel Says Every PC Will Soon Be Capable of Running AI Applications

On Tuesday, speaking at a developer event in San Jose, Intel (INTC) CEO Pat Gelsinger unveiled plans for “the AI PC,” which he says is intended to bring “AI everywhere at scale.”

The company said AI-capable PCs will be enabled by the Intel Core Ultra processor, code-named Meter Lake, which starts shipping on Dec. 14. The new processor will include Intel’蝉 first integrated neutral processing unit “for power-efficient AI acceleration and local inference.”

The battle to provide chips for AI large language model development and inference is intensifying. Nvidia (NVDA) has come to dominate the market for graphics processing units, or GPUs, used in data centers to drive AI model creation, but both Intel and the newly public Arm (ARM) see opportunities for central processing units, or CPUs, that can run AI inference applications on laptops and mobile devices. (Nvidia’蝉 high-end GPUs are too big and too expensive to include in personal computers.)

Having local AI capability on a laptop would shift the direction of the personal computer industry, allowing consumers to run generative AI applications locally, without requiring a connection to the cloud.

Gelsinger says AI on PCs will be a “sea change moment in technological innovation.” Among other things, Gelsinger said the new PCs will be able to run Microsoft’蝉 Copilot software to handle AI tasks for personal productivity applications and other software.

Intel also announced plans to build a large AI supercomputer on Intel Xeon processors and Intel Gaudi2 AI hardware accelerators, with large language model Stability.AI as the anchor customer.

Intel shares were down 2.5% on Tuesday.

Write to Eric J. Savitz at [email protected]