If the United States develops the most innovative semiconductor designs in the world, but can’t manufacture the less complicated chips that its electronics and automobile industries require, is it a case of penny wise, pound foolish?
The ongoing supply chain snafu has made clear the risks in business models that depend on foreign chip production. They were models that made sense before the current dilemma, as evidenced by the profit margins of companies like Intel and AMD.
Now, the U.S. Commerce Department in late January released findings that the American supply of semiconductors is at a critical low. The U.S. must engage in a mammoth effort to replicate the success of Taiwanese and South Korean manufacturers who took a different path where semiconductor production is concerned.
According to Michael Splinter, former Chairman and CEO of Applied Materials, Taiwan and South Korea generally, or Taiwan Semiconductor Manufacturing Company (TSMC) specifically, achieved their current status in the semiconductor industry by investing purely in chip production while American companies focused on product design.
“Early on they might have gotten help from the government, significant help. It wasn’t the most significant factor in how the industry evolved. In TSMC, they came up with a better business model,” Splinter says. “They focused solely on the manufacture of integrated circuits for other people, the so-called ‘pure play’ foundry model.”
Thanks to cheaper labor and fewer regulations, the foundry companies could manufacture chips much more cheaply than smaller integrated device manufacturers, who left the job to the foundry companies, whose growth accelerated, allowing them to invest heavily and long-term in R&D. Now, while the U.S. leads the world in design and product design, for instance, chips intended for AI and machine learning, according to Splinter, its dependence on foreign sources for semiconductors places the U.S. in a bind.
“In one of these semiconductor technologies, there are literally thousands of incredibly intricate steps in the manufacturing process, any one of which could delay or destroy the startup of a new factory. So, I think it’s important that we understand that problem,” Splinter says.
The CHIPS Act is Not A Quick Fix
Splinter made his comments during a webinar held on February 18 by SAFE Commanding Heights, self-described as a “non-partisan initiative dedicated to advancing critical supply chains for America’s transportation and energy needs,” focused largely on support for the CHIPs for America Act, passed by the House of Representatives in early February. It plays a large role in Intel's recent announcement of a $20 billion investment in a new microchip production facility or “fab” in Licking County, Ohio.
According to Sreenivas Ramaswamy, senior advisor to the U.S. Secretary of Commerce, also speaking from the webinar, the government still wants to hear more from customers, investors, universities and entrepreneurs about various aspects of the CHIPs program like workforce considerations, attracting private capital and how to change the fundamental economics of the semiconductor industry.
“We certainly want to solve for the manufacturing capacity and the technology leadership. But can we use this opportunity also to solve for some of these longer-term issues so that we don’t have to deal with these sorts of problems in the future,” said Ramaswamy. If the CHIPs Act is funded, then begins the recruitment of subject matter experts to staff up offices and allow the Commerce Department to begin engaging with a “really complex supply chain.”
This is not a challenge the U.S. can handle entirely on its own. “It’s pretty obvious that there are portions of the supply chain where, if you’re really serious about shoring up capability for the U.S. you’re going to need foreign companies to participate, whether that’s in the tools, in some of the upstream chemicals deals, in the actual fabs and the processes themselves,” Ramaswamy said.
The U.S. will also need to coordinate efforts with allied nations taking their own steps to support the semiconductor industry. “There’s a lot of discussion going on at the government level of how we do this so that there's actually some level of coordination, that it isn't a race to the bottom, that we don’t end up inadvertently over-supplying the industry and then destroying the economics of the industry,” Ramaswamy said.
Splinter adds that China is investing heavily in product innovation, such that the U.S. cannot be complacent about taking steps to defend this lead position while also stepping up manufacturing. More funding and encouragement for R&D to attract Samsung, Intel and TSMC will encourage bleeding-edge semiconductor research in the U.S. More PhDs in the physical sciences will support those efforts. The CHIPs Act supports these measures.
“Intel, they manufacture here and have a system developing new technologies and manufacturing them here. For Samsung and TSMC, that’s not the case,” Splinter says. “Their R&D is done in Korea and Taiwan respectively. And moving these advanced technologies to a factory on a different continent is an incredibly delicate process.”
