We Need More Microchips, So Why Not Build Them Here?

There’s a scene in Raiders of the Lost Ark where Indiana Jones faces off against a guy menacingly brandishing a scimitar. After allowing his enemy to show off some of his moves, Harrison Ford’s Indy character simply pulls out his pistol and shoots him.

That’s how many manufacturing folks feel about the microchip shortage. Instead of trying to maneuver around a problem that threatens to slice and dice them to death, why can’t America just pull out the big ammo – and build our own chip plants? Be done with the problem.

And moreover, wouldn’t Flyover Country be a great place to do that? We have at least five arguments in favor of the next U.S. chip plant being built somewhere between the Appalachians and the Rockies. They come from the arenas of geology, hydrology, logistics, education, and serendipity.

First, to the need for our own plants. Clearly, enough is enough. Rather than losing hundreds of billions of dollars, imperiling jobs and companies, frustrating flush consumers, stoking inflation and reinforcing the disappointing reality that one crucial lever of control of the U.S. economy really resides in Asia, not here – can’t America just build a microchip factory or two?

The answers are “yes” and “it’s complicated.” Constructing new chip plants in this country at the cost of billions of dollars each would take too long to alleviate the current shortages wracking many manufacturing industries. Overall, the chip shortage has cost the car business, for instance, a staggering $110 billion in revenue this year, according to a new forecast by AlixPartners released last week -- nearly double the $61-billion hit forecast by the consultancy just four months ago.

Auto companies and others have been trying ways around the immediate problem, such as using old versions of chips in storage, siphoning chips to their highest-profit models, and juicing software to make up for hardware deficits.

But car CEOs can’t fully compensate for bad decisions they made in 2020 and before, such as getting overly cozy with just-in-time supply chains and, amid the pandemic, cutting production so that the chips they otherwise would have needed were diverted by chip makers to other customers, such as consumer-electronics companies.

“A big part of the problem was that they were too myopic in their views,” Douglas Kent, executive vice president of strategy and alliances at the Association for Supply Chain Management, told me. “There were wild demand swings in chip utilization in non-competing industry sectors during Covid, and they didn’t have a good perspective on that.”

And once they’ve pushed through the current chip shortage that could last several more months, over the long term, manufacturers must adopt new risk-management measures, reconsider inventory practices, and source more generic chips that are easier to source and which they could customize via software to their required specs.

But the most potent solution for U.S. manufacturers and the nation’s economic and military security would be to build more chip plants in this country, including the entire manufacturing supply chain and ecosystem that support them. “We have under-invested in [semiconductor] production and hurt our innovative edge, while other countries have learned from our example and increased their investments in the industry,” Gina Raimondo, U.S. Commerce Secretary, said during a webcast for Intel’s announcement in March. Congress passed a CHIPS for America Act recently, which provides assistance in grants for advanced semiconductor manufacturing and research.

Promisingly, Intel already announced in March that it would construct two chip factories in Arizona, at what is already its largest chip-manufacturing site. Another global giant, Taiwan Semiconductor Manufacturing, has started to build its own $12-billion chip factory north of Phoenix.

Several years ago, Intel abandoned a never-opened chip plant in Colorado Springs, Colo., which then became a site for bitcoin mining – but possibly could still lend itself to repurposing as a chip plant again.

But any real relief for U.S. manufacturers still will be long in coming. The Intel plant, for instance, won’t start production until 2024, the company said. Getting new chip-making capacity up and running requires “a multi-faceted production process that is measured in weeks and months, not in days,” Kent said. “Even in accelerated fashion, an increase in capacity just from the manufacturing-plant and equipment perspective is a multi-year process.”

Given that we’re talking about a long-term solution here, Flyover Country should get due consideration for being part of it. There are five compelling reasons:

Geology: Microchip manufacture isn’t necessarily any more vulnerable to earthquakes than other forms of manufacturing per se, but it happens that 37 percent of the world’s hundreds of chip-fabrication plants are in high-risk areas for earthquake activity, mainly the infamous “Ring of Fire” around the Pacific Ocean that includes China, Singapore, South Korea and California. Another 24 percent is in moderate-risk locations.

Why not put chip plants in the American Midwest or South or Great Plains where, historically, earthquake occurrence and risk are very low? Moreover, Flyover Country for the most part is a region where severe-weather events that also can wreck supply chains – think hurricanes -- have an extremely low incidence. And it would truly be an unlikely twist of fate or perhaps require a precise act of God if some future tornado roaring down a Midwestern wind alley wiped out a new chip plant.

Hydrology: Churning out millions of microchips requires voluminous quantities of water. Right now, the worst drought in half a century is scorching Taipei, Taiwan’s capital, which is home to two-thirds of the world's semiconductor manufacturing capacity. The Taiwanese government so far is making water-allocation exceptions to protect their vital industry, but there are no guarantees that seasonal typhoons later this year will correct the problem.

Meanwhile, Arizona authorities pledge that Intel won’t have any problem getting enough water for its existing and new chip factories. But their promises are imperiled by the fact that the state expects for 2022 to make its first-ever “shortage declaration” about water availability from the Colorado River and that the state simply is due for a drier future because of drought and climate change.

Which brings us to the Great Lakes and even to other Flyover Country states that are aside or astride mighty tributaries such as the Ohio, Missouri and Mississippi Rivers: We have fresh water in abundance for any kind of manufacturing, especially relative to other parts of the country.

Logistics: As the locale of a huge amount of America’s manufacturing activity, including many of the industries that most rely on microchips, Flyover Country is a logical choice for location of chip plants that would optimally shorten supply chains for companies that are used to waiting for their containers of silicon goodies to arrive from halfway around the world.

“When you’re doing network optimization for supply chains,” Kent explained, “typically the betterment in terms of overall costs comes when you can co-locate – even if just regionally – sources of supply and sources of demand. If your primary objective for [chip] capacity is to support the automotive industry, for instance, then Detroit becomes the epicenter” for consideration of where to site a chip plant.

Education: Kent also remarked on the huge advantages enjoyed by much of Flyover Country in higher-educational standards and, more specifically, in the training and employment of engineers, software developers and others who are the ones figuring out how to utilize microchips in their companies’ products.

“There’s a plethora of available workforce and good educational systems in most Midwestern states,” he said, “as well as a lower cost of labor overall, because of the lower cost of living” compared with the coasts.

Serendipity: It just so happens that there’s an existing site in Flyover Country, with a huge sunk investment by a major microelectronics company, which is just waiting to be completed by a huge manufacturing investment and also benefits handsomely from the geologic, hydrologic, logistical and educational benefits listed above.

Foxconn botched its initial pledge to Wisconsin and the region to make huge glass panels for TV screens at its site near Racine, Wisconsin, and now – with scaled-back expectations and greatly reduced financial incentives from the state – is casting about for what to build there. Smaller screens are still a possibility, as are electric cars, under Foxconn’s new deal with Fisker.

But Kathleen Gallagher, of 5 Lakes Institute, has suggested the “compelling idea” of Foxconn building a chip-fab plant on the site. “There are only a handful of fabs in the Midwest,” Gallagher wrote recently in the Milwaukee Journal Sentinel. “There are a lot of reasons our region should have more.”

Foxconn is “not a traditional microchip player and is more of a contract manufacturer, utilizing chip inventories to build other products,” Kent noted. And so unlike, say, Intel, Foxconn likely would need some help in putting together a local ecosystem of suppliers to feed any chip-manufacturing operation.

(This could be a big deal because the difficulty Foxconn perceived of putting together a vast ecosystem of suppliers necessary for making humongous TV screens on the southwestern shore of Lake Michigan was a huge stumbling block in realizing the vision for the site that it initially expressed.)

Some believe that quantum computing may modify or even replace semiconductors someday, Gallagher noted. “But traditional electronics will exist for some time,” helping make the case for the huge investments that would be required in Flyover Country for a chip plant or two.

And once quantum computing takes over, she wrote, our region – not surprisingly – is well equipped to play then, too. Three of the eight new federal quantum-computing centers and institutes are located in Illinois.