Intel CEO Charts Future Path with IDM 2.0, Forms Research Pact with IBM
The semiconductor giant is investing $20 billion to add plants in U.S. and Europe.
March 25, 2021
In a major pivot, Intel CEO Pat Gelsinger says the semiconductor giant will manufacture chips for competitors. Intel will also invest $20 billion to aggressively add manufacturing plants in the U.S. and Europe.
Looking to the future, Intel is partnering with IBM.
Gelsinger this week delivered the sweeping and aggressive blueprint, describing how he intends to restore Intel’s technical and market leadership.
As the largest provider of semiconductors for PCs and servers, Intel has fallen behind its rivals in recent years. Looking to reverse that tide, Intel tapped Gelsinger, as CEO in January. Gelsinger’s presentation represents swift actions that could shape the competitive environment for device, data center and cloud compute infrastructure for years to come.
Intel’s Pat Gelsinger
“We are the only company with the depth and breadth of software silicon platforms packaging and process with at-scale manufacturing our customers depend on for their next-generation innovations,” the Intel CEO said. “To meet this moment and position our company for the future, I am setting a course for a new era of innovation and technology leadership.”
Besides offering Intel’s client and data center CPU road map, Gelsinger revealed an ambitious plan to create more efficient manufacturing capacity. Intel describes the effort as the second generation of its integrated device manufacturing (IDM 2.0) model.
Intel’s $20 billion expansion of manufacturing capacity in the U.S. and Europe aims to lessen its dependency on one region. Currently, 80% of Intel’s manufacturing plants are in Asia. The company broke ground on the first of two new plants on Monday at its campus in the Phoenix area. Gelsinger said Intel will launch a second site in the U.S. and one in Europe within the year.
Intel Foundry Services
The most notable departure for Intel is the decision to provide manufacturing capabilities to other chipmakers, including competitors. The company launch of Intel Foundry Services is a separate business unit. Gelsinger said Intel Foundry Services will satisfy unmet and growing demand for semiconductors.
“We see strong market demand overall,” the Intel CEO said. “The combination of our internal and external capacity requirements allows us to build ahead, create more capacity, that we can be satisfying those collective needs of our customers. And underneath that we’ll meet our commitments.”
As a separate business unit, Intel Foundry Services will be accountable for its own profit and loss performance, Gelsinger said.
“Obviously, we’re going to work hard to always be on the front foot of the technology, of the capacity and overachieving on every aspect of our businesses as we look forward,” he said.
Historically, Intel has shunned such a move, despite frequent calls for …
… the company to do so. Industry analyst Jack Gold of J. Gold Associates said the creation of Intel Foundry Services is a controversial decision.
J. Gold Associates’ Jack Gold
“Outsourced manufacturing may be attractive from a ‘bookkeeping’ position but is not necessarily the best way to build high-performance processors,” Gold wrote in a research note.
Despite Gelsinger’s promise of potential “coopetition,” shares of rival companies that manufacture chips, including Apple, Qualcomm, Samsung and TSMC, fell on Tuesday.
Partnership with IBM
Intel has also tapped IBM to collaborate on research to advance silicon process and packaging technology.
“Both are foundational elements of manufacturing leadership, and Intel and IBM have been incredible sources of related innovation over the last 30 years,” Gelsinger said.
“This represents an important milestone in the world of technology,” said IBM CEO Arvind Krishna, who made a brief appearance during the webcast. “By bringing together two of the best semiconductor research organizations in the world, we will greatly accelerate innovations in the semiconductor industry.”
Krishna endorsed the launch of Intel Foundry Services and Intel’s decision to add manufacturing plants in the U.S. and Europe.
IBM’s Arvind Krishna
“That will bolster the United States competitiveness in semiconductors,” Krishna said. “We also understand the importance of developing a more secure supply chain. It’s a vital importance for our economy, and for national security.”
Roger Kay, principal analyst of Endpoint Technologies Associates, described Gelsinger’s one-hour presentation as “one of the most momentous speeches I’ve heard in my decades in the technology industry. It was so comprehensive and far-ranging that it will literally set the pace for peers, rivals, customers and suppliers for years to come.”
In another departure for the company, Gelsinger said Intel will expand its use of third-party foundries, notably with TSMC.
Nevertheless, Gelsinger said Intel will maintain the bulk of its manufacturing among its own plants. The company’s delayed move to 7nm process node will go live next quarter for its Meteor Lake processors. Intel is using extreme ultraviolet lithography (EUV), which said it will provide a more simplified process flow.
Intel’s Processor Road Map
Intel made official that it will also launch its 3rd Gen Xeon Scalable processors, code-named “Ice Lake,” on April 6. It would come just weeks after rival AMD launched its Gen 3 EPYC datacenter processor.
“With the Ice Lake platform and our data center portfolio, we deliver incredible flexibility for our customers and continue our vision of delivering AI everywhere from edge to the cloud,” Gelsinger said.
Gelsinger said customers are also now testing Intel’s next data center processor — Sapphire Rapids. The goal is to start production by the end of this year, eyeing a mid-2022 rollout.
On the road map for 2023 is the “Meteor Lake” processor for PCs and client devices, and Grand Rapids for data center infrastructure. Both Meteor Lake and Grand Rapids will have compute tiles built on Intel 7 nanometers, Gelsinger said.
“And the strength of our new IDM 2.0 model means we can strategically use the breadth of our ecosystem to our advantage for 2023 road map,” he added.
Intel’s long-delayed 7 nanometer process manufacturing will commence next quarter with its Meteor Lake processors. The 7nm shift is “a pivotal aspect of how we’re engineering the future,” Gelsinger emphasized.
Reluctant to embrace extreme ultraviolet lithography (EUV) for chip printing resulted in its 10nm manufacturing processors later than others. That had a cascading effect on the transition to 7nm, leaving Intel behind in the use of EUV.
“We’ve re-architected and simplified our 7nm process flow, increasing our use of EUV by more than 100%,” Gelsinger said.
Read more about:
VARs/SIsAbout the Author
You May Also Like