4 September 2022

Can Semiconductor Reshoring Prime a U.S. Manufacturing Renaissance?

Sujai Shivakumar, Charles Wessner

Recognizing the strategic importance of renewing U.S. innovation and manufacturing, the Biden administration in 2021 launched an initiative to foster the “reshoring” of several key industrial sectors, particularly semiconductor manufacturing. Despite the bipartisan realization that the United States is now lagging compared to other countries’ production capabilities in semiconductors both in quantity and—crucially—in quality, an agreement on related legislation was slow to materialize. Fortunately, such an agreement finally came in August 2022 in the form of a sweeping $280 billion bill that provides support for scientific research and includes an unprecedented $52 billion in federal funding to promote the expansion of chipmaking within the United States. This new strategic focus on semiconductor manufacturing represents a more comprehensive understanding of the industry’s importance beyond a secure supply of chips for defense applications.

CHIPS as a Response to Supply Chain Shortages

The Covid-19 pandemic heightened public and congressional awareness of the importance of semiconductors for civilian applications. as shortages of chips began to hamper production of automobiles, smartphones, and other key consumer products. However, the pandemic was merely a disruptive catalyst that exposed long-running U.S. vulnerabilities in strategic supply of semiconductors. U.S. onshore chip manufacturing capacity had been eroding for decades as the United States became increasingly reliant on extended and fragile global supply chains, especially for the most advanced chips. The Russian invasion of Ukraine further underscored these vulnerabilities just as dependency on Taiwan (TSMC) and South Korea (Samsung) for semiconductor chips reached new heights.

Encouraged by the prospect of the CHIPS Act and its powerful incentives, an unprecedented number of new semiconductor wafer fabrication plants are now under construction or planned in the United States. These reflect major investment plans by both U.S.- and foreign-owned companies such as Intel, Texas Instruments, Micron, Taiwan’s TSMC, South Korea’s Samsung, and GlobalFoundries, among others. If and when they become operational, the new facilities will reduce U.S. dependency on foreign chipmaking operations and benefit all downstream U.S. manufacturing industries.

CHIPS as a Catalyst for Innovation Clusters

As a result of the “spillover” benefits associated with these unique high-tech investments, the establishment of new semiconductor manufacturing plants in the United States is likely to stimulate domestic manufacturing across a broad spectrum of industries.

Economic development professionals recognize that chip plants are commonly the focal points of manufacturing and innovation “clusters,” characterized by concentrations of technology-intensive companies, large numbers of skilled workers, and academic institutions and training centers offering curricula relevant to chip design and production. It is worth examining the elements of this “clustering” to appreciate how new chipmaking facilities are likely to catalyze additional manufacturing activity in the U.S. regions where they are located.
Ramping Up Nearby Manufacturing Operations

Building advanced semiconductor fabs can draw investments to nearby manufacturing operations that provide the equipment and materials needed for those fabs to operate. According to John VerWey, East Asia national security advisor at the Pacific Northwest National Laboratory, “When you build a leading-edge fabrication facility, you’re not just getting a leading-edge fabrication facility. You’re getting the ecosystem of chemical suppliers, material suppliers, mask suppliers, equipment suppliers that fill that factory.” At present, plans by Intel, TSMC, Samsung, GlobalFoundries and other chip firms to establish new fabs in the United States are attracting billions of investment dollars into U.S.-based manufacturing facilities not just for chips, but also for the chemicals and materials needed for domestic chip fabrication. New innovations are a major driver, as “the new fabs will be hungry for chemical raw materials because of the growing number of processes carried out on each wafer . . . It’s the number of lithography steps, the number of cleaning steps, the number of deposition steps, the number of etches,” according to Michael Corbett, a professor at the Tufts University School of Engineering. As the steps increase, the need for materials, equipment, supplies, and talent to manage these incredibly high-tech processes grows as well.

The state of Texas is a powerful example. It has over 200 semiconductor-related manufacturing facilities, the vast majority of which consists of equipment manufacturers and material suppliers for local device makers like Texas Instruments, NXP, Infineon, and Samsung. These upstream firms expand as the chipmaking plants grow and proliferate.

This was illustrated in August 2021, when Samsung was considering a new fab in central Texas. Texas-based Schunk Xycarb Technology Inc., a manufacturer of quartzware used in semiconductor production, announced that if Samsung chose Texas, Schunk would build a second facility there and “expedite hiring and equipment purchases.” Likewise, Huntsman Corporation announced in September 2021 that it will expand capacity in Texas for specialty semiconductor-grade amines to support “the growing and dynamic semiconductor industry” and to “satisfy many of our customers’ requests to have a North American supply source for these vitally needed products.”

Similarly, Merck KGaA said in December 2021 that it would spend about $1 billion on electronic materials, research and development (R&D), and capacity expansion at four locations in the United States to support expanded domestic semiconductor production. In January 2022, Taiwan’s Sunlit Chemical began construction of a $100 million plant in Phoenix—where TSMC is building a wafer fab—to make hydrofluoric acid, which is used in semiconductor etching and cleaning. Meanwhile, Air Liquide announced that it would invest $60 million in onsite plants in Phoenix to supply ultra-high purity hydrogen, helium, and carbon dioxide to a local semiconductor manufacturer.

Creating Offshoot Industries

In addition to ramping up the extraordinary technological and industrial competencies required to manufacture advanced semiconductors, new fab construction can support the creation of offshoot industries that capitalize on those competencies.

In collaboration with the Dutch firm QuTech, Intel plans to repurpose conventional silicon manufacturing processes to manufacture silicon qubits, which are needed in huge quantities to support quantum computing. “Our research proves that a full-scale quantum computer is not only achievable but also could be produced in a present-day chip factory,” Intel’s director of quantum hardware, James Clarke, claimed.

Furthermore, Austin-based Silicon Labs, a designer and manufacturer of semiconductors, announced in 2021 that it would sell its automotive and infrastructure businesses to focus entirely on the technologies needed to enable the internet of things (IoT), a term used to describe the connection of non-computing objects such as clothing, wall covering, kitchen appliances, and packaging to the internet. Semiconductor technology is central to the IoT, which will involve the incorporation of transistors and electronic circuits on substrates like paper, plastic film, metal foil, wallpaper, and cloth fabric. Although it is too soon to know whether Silicon Labs and other firms like it will succeed or fail, its core competencies in semiconductor technology represent an enormous asset. Observers predict that the IoT—linked with artificial intelligence—will lead to revolutionary advances in productivity, convenience, efficiency, and innovation in many areas of human endeavors.

Growing the Region’s Skilled Workforce

Semiconductor manufacturing relies on the local presence of substantial physical and human capital, including, most importantly, a workforce that possesses the skills, discipline, and stamina to engage in high-tech manufacturing. A deep talent pool and enabling educational institutions (both research universities and specialized community colleges) to supply training and physical infrastructure is a powerful element in attracting semiconductor manufacturing. Moreover, these same institutions and facilities can attract investment in other types of advanced manufacturing with similar support requirements.

In this regard, when Intel announced plans in 2021 to spend $20 billion to build two new fabs in Arizona—where other fabs were already located—a company spokesman observed that “clustering of multiple fabs in one location offers many benefits, including access to a robust talent pipeline, an established ecosystem of supply chain partners, and reliable infrastructure to support our electricity, water, and other utility requirements.”

For Austin, Texas, touting the “region’s unique workforce is one of the main things attracting new factories” from firms as well as other manufacturers, including Tesla, Plastikon Industries, Continental AG, and others. The founder of the Austin Regional Manufacturers Association (ARMA), which has focused heavily on workforce development, remarked in 2021, “What’s unique about Austin compared to the rest of Texas is our advanced manufacturing workforce. There’s a lot of electronics, instrumentation and semiconductors being built here, which has created a dynamic workforce that is appealing to many of these advanced factories that are coming here, like Tesla.”

Fostering Cross-Industry Collaborations

Recent events have underscored how the presence of a semiconductor manufacturing plant in a locality can attract manufacturers from other tech-intensive sectors who seek research and production collaborations with a chipmaker.

Tesla moved its headquarters from California to Texas in 2021 and began production of Model Y electric vehicles (EVs) at a new factory on the outskirts of Austin a year later. In part, these moves reflect its partnership with Samsung, which is concurrently expanding its chipmaking capacity in Austin. Samsung already manufactures Tesla’s full self-driving (FSD) chips at its fab in Austin, and Samsung LEDs are being used in Tesla “smart headlamps.” Samsung’s new fab in Texas—40 minutes away from Tesla’s factory—is expected to produce next-generation FSD’s for vehicles like the Cybertruck, which will feature Tesla’s Hardware 4.0 computer. As the top EV manufacturer in the world, its partnership with Samsung likely began when it was searching for electronic parts necessary for making EVs and autonomous vehicles.

In 2018, AT&T and Samsung jointly announced the creation of the nation’s “first manufacturing-focused 5G Innovation Zone” on Samsung’s Austin campus, which now features AT&T’s wireless technology along with Samsung’s 5G network equipment. The companies said they chose Austin as the site “because of its strong semiconductor manufacturing industry.” The zone’s purpose is to “provide a real-world understanding of how 5G can impact manufacturing and provide insight into the future of a Smart Factory.”
Encouraging Spin-Offs

Since the inception of the semiconductor industry, chip manufacturers have spun off new local enterprises founded by company alumni, a practice which has continued to the present day. Spinoffs occur as employees of the company leave to pursue new ideas and strategies on their own, or if the parent company decides to enter a new field or divest itself of certain lines of specialization. While many of these ventures fail, some succeed in spectacular fashion—in some cases completely eclipsing the parent company.

The most famous example of this phenomenon is Fairchild Semiconductor, founded in 1957 in Mountain View, California, in what later became known as “Silicon Valley.” Fairchild alumni founded Applied Materials, Intel Corporation, National Semiconductor, and Advanced Micro Devices, which pioneered some of the most important innovations in the industry's history, including integrated circuits and MOS and CMOS technologies. In 2014, a research study traced over 90 Bay Area tech companies to the founders and employees of Fairchild, with a market value at that time of over $2 trillion.

In 2008, a subsidiary of Taiwan semiconductor maker UMC, NextPower Technology Inc., was the first Taiwanese company to enter high-volume production of thin film solar cells. In 2010, Taiwanese chipmaker TSMC invested $258 million to create the Thin Film Solar R&D Center in Taiwan. In 2011, TSMC spun off its solar business, by which time Taiwan’s solar cell industry ranked number two worldwide, surpassing Japan and Germany's industries.

In 2003, TSMC created VisEra as a spinoff to develop and operate foundry production of optical sensors. In 2016 VisEra became a wholly owned TSMC subsidiary. In 2021, TSMC—by then the sole owner—began divesting its shareholdings in VisEra, which was going public. Today VisEra is the world’s leading independent image sensor foundry provider and is the only foundry capable of manufacturing ultra-thin optical fingerprint solutions.

Advantages of Cross-Pollination

Finally, as those familiar with studies of industry clusters are aware, the concentration of technology-intensive companies in a given geographic area has a “cross-pollination” effect as creative individuals from various backgrounds convene formally or informally to share ideas and information or move from job to job within a cluster. As Alfred Marshall, the nineteenth-century British economist who made the first academic studies of industry clusters put it, in such localities, “the secrets of the trade are in the air.” Such benefits are serendipitous and cannot be readily quantified or measured, but they are real and observable. The result can be the creation of new technologies, new companies, and entirely new industries.

In his highly acclaimed book The New Economy of Jobs, Berkeley professor Enrico Moretti offers numerous real-world examples of the cross-fertilization that can occur when different high tech industrial communities are co-located. He observes, for example, that in Silicon Valley the existence of two seemingly unrelated sectors, medical research and video gaming, have “become intertwined in the form of ‘serious games,’ products that apply cutting-edge gaming technologies to cure diseases,” such as “game-like software intended to improve memory and attention and possibly even help treat such disorders as autism and schizophrenia.”

These effects can be seen in Hsinchu Science Park, where Taiwan’s semiconductor industry was established. The result is a dense collection of research organizations, manufacturers, and supply chain firms where myriad interactions between individuals working in the park in related sectors are a principal reason why it “is arguably the pivot on which Taiwan’s modernization turned [and in which firms such as UMC and TSMC] transformed the island into a key center of the global electronics industry.” In a Stanford Institute for Economic Policy Research report, the authors observed that “the intense communications within this community fosters imitation, joint problem solving and transfer of information and know-how about management, technology, the job markets, and new firms and products.”
Putting It Together

It is important to note Taiwan and Austin are not the only examples positive dynamic effects of semiconductor investment made possible through active public and private partnerships. New York state’s “regional renaissance” is an instructive case study of how public and private initiatives can work together to develop a vibrant nanocluster. The unique research facilities at the College of Nanoscale Science at SUNY-Albany became a key factor behind the decision of Global Foundries to build a fab in Upstate New York. New York had to compete with places like Austin, Dresden, and Singapore, but existing research facilities, combined with New York’s excellent educational system and backed by substantial state subsidies, succeeded in attracting a major foundry (a contract fab) to the region. The state’s investment was a huge success, creating three times the expected direct jobs, substantial additional jobs along the supply chain, and providing a massive stimulus to regional growth—and the associated tax receipts. GlobalFoundries then expanded without state aid. Now, aided by the CHIPS Act, there are plans to further expand, which will again create huge numbers of construction incomes and additional high-tech, high-paying jobs across the region.

As described above, it is important to keep in mind that investments in semiconductor manufacturing create new opportunities in related high-tech industries and, perhaps more importantly, in key emerging industries such as artificial intelligence, quantum computing, and IoT. The presence of advanced research facilities, combined with hands-on manufacturing knowhow are keys to future regional and national growth.

Powering a Manufacturing Resurgence

The specific benefits from new industries that will accrue to a given region are difficult to predict. However, given the straightforward and powerful stimulative effects of chip manufacturing on the broader manufacturing sector within a region, there is every reason to believe that the numerous new and planned chip fabs in Texas, New York, Arizona, Ohio, and other states could help power a broader resurgence in U.S. manufacturing with its attendant benefits in employment, skills, technological competency, and ultimately, U.S. national security.

There is an important caveat: if the nation is to reap these benefits, the federal government and the states will need to cooperate in both incentivizing and facilitating the private sector’s ability to move forward expeditiously with new construction projects and accelerate the environmental reviews and local permitting processes that go with it. This will take cooperation across state and local governments and industry to rapidly create win-win solutions. CHIPS represents a major opportunity, but shrewd implementation is crucial if this opportunity is to bear fruit.

No comments: