Verified, then logged.
Sun 22 Nov 2026 · 00:23 KST / 15:23 UTC한국어 | ENSearch /

08 / Industry · Analysis

Intel sets out separate CPU and GPU roles for AI computers

A division of processor roles linked memory sharing with the choice of where to run AI tasks.

Wafer processing tools line an amber-lit memory chip cleanroom as wafer carriers travel along ceiling rails and a technician in a cleanroom suit stands blurred in the distance with their back turned. AI-generated image

Intel presented a strategy for dividing work between central processing units (CPUs) and graphics processing units (GPUs) at AI Summit Seoul 2026, according to TechM, a Korean technology publication.

TechM framed the strategy around a changing role for computers as AI moves beyond answering questions towards carrying out tasks itself. That was the report’s interpretation of the shift discussed at the event.

The event also covered memory sharing and choosing an AI execution environment to suit the task, according to TechM. Together, these presentations connected the allocation of processor work with how memory is shared and where AI runs.

Dividing work between processors

According to TechM, Intel presented heterogeneous computing as a strategy that assigns different roles to CPUs and GPUs. Heterogeneous computing means using different types of processors together, with work divided between them.

SK hynix, the Korean memory chipmaker, introduced memory-sharing technology, TechM reported. The report treated shared memory alongside the division of processor roles as part of the changing design of AI computers.

Choosing where AI tasks run

Lenovo, the computer maker, introduced a hybrid AI strategy that selects an execution environment according to the task, TechM reported. In that account, the work to be performed provided the basis for choosing where AI runs.

Intel also pursued a broader AI ecosystem through cooperation with 23 Korean companies, according to TechM. The report thus connected the division of work inside computers with choices about AI execution and cooperation between companies—the elements presented for computers taking on more tasks.