Intel to Make AI GPUs Says CEO Lip-Bu Tan
By Reuters | 03 Feb, 2026
The plan to enter a segment created by Nvidia reflects both Intel's state-of-the-art chip fabrication foundry and Nvidia's struggle to meet overwhelming demand for the data center chipsets used to train AI models and power their inference work.
Intel CEO Lip-Bu Tan on Tuesday said the company plans to build graphics processing units (GPUs), the category of chip popularized by Nvidia.
"I just hired the chief GPU architect, and he's very good. I'm very delighted he joined me," Tan said, claiming that it took some persuading.
Qualcomm executive Eric Demmers last month went to Intel, a move first reported by industry publication CRN and later confirmed by Demmers on LinkedIn.
In an interview with Reuters on the sidelines of the Cisco AI Summit, Tan said the GPU effort will target data centers, where Nvidia has built a massive business in recent years, and that Demmers will report to Intel's data center chip chief Kevork Kechichian.
"It's tied in with the data center," Tan told Reuters. "We're working with customers, and will then define what the customer needs."
Tan said from the stage that "a couple of customers are engaging heavily" with Intel's chip contract manufacturing operation, called Intel Foundry. In the interview with Reuters, he said that the interest was around Intel's 14A manufacturing technology and that volume manufacturing would likely ramp up later this year.
"In order to have a customer ... they have to let us know what is the volume and which product, so that we can plan and take time to build the capacity," Tan told Reuters.
Tan also said that during a recent hiring drive for chip designers, he was "shocked" to find that Huawei Technologies Co had hired about 100 "top notch" designers, despite the fact that the U.S. has restricted its access to chip industry software and tools.
Tan said that when he asked Huawei designers why they had joined the Chinese firm when it does not have access to U.S. tools, "they said, 'Even though we don't have access to the best tools, like (electronic design automation) tools from Cadence and Synopsys, we have the poor man way to do it, and we can do it,'" Tan said.
"To me, they are just shortly behind us, and if you're not careful, they will just leap forward ahead of us."
(Reporting by Jeffrey Dastin in San Francisco,; Writing by Stephen Nellis; Editing by Mark Porter and Nick Zieminski)
Recent Articles
- Oracle Workforce Shrinks by About 21,000 Employees Amid AI Adoption
- Ohtani’s Baby Boom, Kim’s Birthday Celebration, and World Cup Updates
- Become an AI Bonus Baby—or Just Get Paid Like One
- Ferrari Denies Requiring Luce EV Purchase to Access Limited Edition Models
- SpaceX Turns to Bond Sale to Fatten $100.8 Billion Cash Stash
- Google DeepMind Signs AI Research Deal with Film Studio A24
- China's 618 Shopping Festival Sees Flat E-Commerce Sales from Cautious Shoppers
- Amazon Prime Day to Gauge US Consumer Strain as Focus Shifts to Basics
- China Closing in but US Still Leads in Biotech Innovation
- EV Surge Likely to Cut Oil Demand by Late 2027
