After decades of leading the PC CPU space, Intel is putting more weight behind its enterprise and data center business, a move that’s starting to shape its plans in the GPU arena as well.
Intel CEO Lip-Bu Tan announced at the Cisco AI Summit that the company has named a new chief GPU architect. Although the executive was not identified during the event, subsequent reports revealed the role went to Eric Demmers, previously of Qualcomm, who Tan said was persuaded to join Intel.
Joining Intel last month, Demmers reports directly to EVP Kevork Kechichian, head of the Data Center and AI Group. The alignment highlights Intel’s intent to prioritize GPU development for data center and AI workloads over consumer-focused gaming products.
The appointment helps address a gap in Intel’s product portfolio. While Arc GPUs serve the consumer market, Intel does not yet have a competitive solution aimed squarely at AI training and inference, a responsibility expected to fall under Demmers.
The decision is consistent with Intel’s stated manufacturing priorities. Earlier this year, the company said that high-performance Xeon processors for enterprise and workstation customers now account for most of its production capacity, pushing consumer Core CPUs to a lower priority.
Intel has already begun executing its enterprise GPU. The company unveiled its Crescent Island data-center GPU architecture in October, designed for AI workloads and built on the broader Xe platform, which emphasizes higher memory capacity and energy efficiency. Intel is also investing in agentic AI, heterogeneous systems, and AI-enabled PCs.
The company continues to face challenges in this market. Gaudi 3, Intel’s most recent dedicated AI accelerator launched in April 2024, did not reach its $500 million revenue target by year’s end.
Whether Demmers will oversee Crescent Island or a future GPU architecture remains uncertain. At the Cisco event, Tan emphasized that Intel’s priorities are aligned with broader industry trends, citing unprecedented AI-driven demand for compute and memory and predicting that supply constraints will persist through 2028.
Hiring a high-profile chief GPU architect signals Intel’s push to become a credible player in enterprise AI silicon, despite market challenges and a shift away from its consumer PC focus.
Maybe you would like other interesting articles?