Recently appointed Intel CEO Pat Gelsinger Tuesday held his first major briefing regarding his plans to bring the company back from several years of challenges.
In an event titled “Intel Unleashed,” Gelsinger spoke on a live video feed broadcast on the company’s news home page of its Web site. Among the major items, Gelsinger said Intel had rectified Intel’s production issues in making 7-nanometer chips; said the company will spend $20 billiion making two factories in Arizona, as part of a plan to expand Intel’s U.S. and European manufacturing facilities; and said Intel will become a leading global foundry to make chips for other companies, including its competitors. Gelsinger also announced Intel is on course to exceed its financial outlook offered in January.
Gelsinger started off by reiterating his view that running the comapny, where he spent decades, is his “dream job,” and that Intel’s “best days are ahead.” Gelsinger, who worked at the company for thirty years before becoming head of software maker VMware, recalled the culture of Intel under founders Gordon Moore, Bob Noyce, and Andy Grove, and said the company’s culture was coming back, speaking of a new excitement among staff. “The old Intel is the new Intel,” he said.
Moving on to discussing the business, Gelsinger declared, “Today, we are unleashing the iconic Intel innovation engine.”
He outlined four objectives:
- Be the leader in every cateogry in which we compete
- Excute flawlessly to our commitments.
- Innovate passionately with boldness
- Foster a vibrant culture
Gelsinger gave an update on the comapny’s 7-nanometer production, which has been repeatedly delayed in recent years. Gelsinger said Intel had in past limited its use of “EUV,” extreme ultra-violet lithography, the means of making the finest chip features. That put the company “on the wrong side of the EUV maturity curve,” he said.
Intel has rectified that, he said, and has a “very strong partnership with ASML,” the vendor that produces EUV.
“I’m pleased to share that we have now fully embraced EUV,” said Gelsinger. “We’ve re-architected and simplified our 7-nanometer process flow, increasing our use of EUV by more than 100%.”
“We expect to tape in our 7-nanometer compute tile for Meteor Lake later this year,” said Gelsinger, referring to Intel’s planned 2023 client CPU.
Gelsinger transitioned to talking abou the company’s forthcoming GPU chip, Ponte Vecchio, which he called an AI supercomputer.
The main event, said Gelsinger, was “The next major evolution of Intel’s IDM model, IDM 2.0.”
Gelsinger called it a “leadership combination of three vectors.” Vector one will be to make the majority of Intel’s own products in the company’s own factories, as Gelsinger had said back in January.
Second, Intel will also make use of foundries outside of Intel. “We are increasing our engagment with TSM, Samsung and Global foundries, and UMC” said Gelsinger.
Third, Gelsinger said Intel will become a “world-class foundry business,” an opportunity he called a $100 billion annual opportunity by 2025. “Intel Foundry Services” will be a new business division reporting to Gelsinger, he said.
Intel is also bidding for U.S. Department of Defense contracts, he said, and is going to build a secure faciity to manufacture parts that impact national security. “We are honored to be competing right now for the U.S. Department of defense contract to develop a domestic commercial foundry that will also meet the security needs of the U.S. government,” said Gelsinger.
Intel said the company had already received support from numerous companies, including Amazon and Google.
Gelsinger introduced Microsoft CEO Satya Nadella, who appeared in a video feed. Nadella said the companies were entering “a complete new era” with “radical changes in compute architecture.”
“We will need to collaborate to co-design” said Nadella. “We applaud your continued investments … in offering new options for silicon designers that will increase manufacturing here in the United States.”
Gelsinger said Intel will spend $20 billion to build two new factories in Arizona, producing 3,000 new jobs.
Gelsinger said the company will focus on making “PDK models” that enable easy transfer of chip designs from Intel’s factories to outside factories.
Gelsinger announced a partnership for chip design with IBM, and he invited Arvind Krishna, IBM’s CEO, in another video feed.
“Our IDM 2.0 plans will put us on a path to restore process-performance leadership,” said Gelsinger.
Turning to talking about products, Gelsinger said the company “remains committed to a steady cadence of leadership products” using both its own and external manufacturing.
Focusing on CPU, Gelsinger said the company’s “Tiger Lake” part continues a “strong ramp,” and he referred to “Alder Lake,” a part coming later this year.
Then Ice Lake will bring AI capabiltiles “from the edge to the cloud,” said Gelsinger. A data center part called “Sapphire Rapids” will arrive in early 2022. Two 2023 parts, Metor Lake and Grand Rapids, will use “tiles,” he said, and will make use of TSMC’s production capabilities.
“For Q1, we expect to exceed our guidance for both top and bottom line,” said Gelsinger. The company has issued an 8-K release with the SEC with the details.
Gelsinger recalled the company’s annual events, the Intel Developer Forum, or IDF. “I’m thrilled to announce we are bringing back the spirit of IDF with a new event series, ‘Intel ON,’ taking place this October in San Francisco.”
Wrapping up, Gelsinger said, “Today is the the first step, and we are just getting started.”