Pat Gelsinger, CEO of Intel Corporation, presented a COMPUTEX Keynote speech on Tuesday that showcased how the company’s next-generation data center and client computing products are making AI technologies easily accessible through exceptionally engineered platforms, secure solutions and open ecosystems support.
Speaking at the Taipei Nangang Exhibition Center on day one of the COMPUTEX trade show, Gelsinger set out Intel’s unique position as a company servicing the entirety of the AI continuum and enabling AI everywhere, from data centers to networks, PCs, cloud and edge computing.
Unveiling the latest offering in the company’s Xeon family of processors, Gelsinger explained that the new Xeon 6 with E-Cores features a high core count and exceptional performance per watt, meaning increased performance for less power. With more than 130 million Xeons already being employed around the world, this new generation will be “an essential upgrade for the modern data center”.
A performance-focussed version of the Xeon 6 with P-Cores, codenamed Granite Rapids, is also slated for later this year.
Gelsinger also showed off Intel’s Gaudi series of AI accelerators, which are offering customers open and cost-effective solutions for powering AI, with an inventive demonstration of their ability to drive a multimodal LLM. The model instantly analyzed a chest X-ray, before comparing Gaudi’s pricing to that of its competition. At $65K and $125K, Gaudi 2 and 3 were found to be one third and two-thirds of the cost as compared to the H100.
Intel view the advent of AI as a turning point comparable to the development of WiFi 25 years ago, whose rapid and universal uptake quickly made it a necessity. By 2028, 80% of all PCs are expected be AI PCs, Gelsinger said, and the company’s Core Ultra devices with built-in AI acceleration have been leading the way for this shift, with 40 million units shipped by the end of the year, 8 million AI PCs shipped to date, and more than 500 AI models already optimized for Core Ultra.
Satya Nadella, CEO of Microsoft, explained in a partner video that Intel’s new Lunar Lake chip, set for launch in Q3, has been essential for the rollout of its Copilot+ AI PCs, by powering more than 80 new laptops from 20 OEMs. These devices “will deliver exceptional security and extended battery life, with a 40-plus TOPS NPU unlocking capabilities simply not possible on other PCs”.
The Lunar Lake architecture delivers an increase of up to 50% better GPU performance, 4X NPU AI compute, 40% lower SoC power, and 3.5X GPU AI compute as compared to the previous generation. “Simply put: Lunar Lake is the flagship platform for AI PC innovation bar none”, Gelsinger concluded.
