Intel Announces Major Return to GPU Development Under New CEO

Intel Makes Bold GPU Play

Intel Corporation has announced plans to develop graphics processing units, marking a significant strategic shift for the chipmaker as it seeks to compete in the booming artificial intelligence hardware market. CEO Lip-Bu Tan revealed the news at Cisco's AI Summit in San Francisco, telling attendees that Intel has already hired a chief GPU architect to spearhead the initiative.

New Leadership Driving Change

Tan, who became Intel's CEO in March twenty twenty-five, framed the GPU effort as a top executive priority. He told the audience it "takes some persuasion" to recruit top architectural talent, underscoring the competitive battle for chip design expertise. While Intel has marketed graphics products in recent years, the emphasis on hiring a dedicated chief architect signals that GPUs are now central to the company's turnaround strategy.

AI Inference Focus

Intel's GPU ambitions include a data centre processor codenamed Crescent Island, announced in October twenty twenty-five. Built on the Xe three P Celestial architecture, the inference-optimised accelerator will feature one hundred and sixty gigabytes of memory and is designed for air-cooled enterprise servers. Customer sampling is expected in the second half of twenty twenty-six.

Manufacturing and Materials Strategy

Tan also highlighted Intel's manufacturing progress, saying the company is "doubling down on glass" as a key chipmaking material and that its eighteen A process technology is performing well. Microsoft and Amazon have both committed to using eighteen A for next-generation AI accelerators. The eighteen A node powers Intel's recently launched Core Ultra Series three processors, the first two-nanometre class chips designed and manufactured in the United States.

Competitive Pressure and Geopolitical Context

The CEO warned about US competitiveness in open-source AI relative to China, expressing shock at reports of aggressive hiring of chip architects by Chinese firms. The comment reflects growing concern that talent competition has become intertwined with the AI hardware race, as companies worldwide race to develop next-generation accelerators.