Sections

Commentary

The chip shortage won’t be fixed without major federal investment

Cover photo

Intel’s recent announcement that it plans to build a $20 billion semiconductor plant outside Columbus, Ohio unleashed euphoric headlines, with President Joe Biden calling it a “game changer” that would begin to revive the Midwest’s industrial sector and address supply chain bottlenecks that have led to the current computer chip shortage.   

But for all the excitement, it bears noting that Intel’s plans are the latest development in a long and often painful saga that is now challenging the U.S. to work out a solution to the nation’s broader chip problems. Addressing those issues will require urgent action, including passage of the pending America COMPETES Act and its important CHIPS for America Fund. 

Semiconductor devices matter because they power and control virtually the entirety of today’s digital economy, ranging from cars and computers to artificial intelligence. As such, semiconductors have become an essential general purpose technology, which means that their continued availability and improvement through reliable supply chains and innovation will be essential to developing modern technologies that improve lives. 

The problem is that the nation’s semiconductor industry has again become fragile. The U.S. created the industry in the 1960s, but lost global leadership in it in the 1970s. After regaining footing in the 1980s, the nation is again now watching the industry erode due to “inattentiveness,” as the Information Technology and Innovation Foundation’s (ITIF) Stephen Ezell has written 

As Ezell showed, stability and continuous innovation in advanced technology products require constant vigilance and a strong industrial strategy, given constant competition from other nations. In the 1980s, for instance, the U.S. responded to Japan’s aggressive push for semiconductor dominance with new investments in research and development, as well as initiatives such as SEMATECH, a public-private research consortium that provided nearly $900 million in federal funding to semiconductor firms between 1986 and 1996. These moves—along with the embrace of a “fabless” production model, by which firms emphasized innovative design and outsourced much of their manufacturing capacity to Taiwan, South Korea, and, increasingly, China—brought about at least a temporary recovery of the U.S. semiconductor industry in the 1990s. 

Since then, however, the longer-term consequences of offshoring combined with aggressive foreign competition have again degraded the U.S. semiconductor industry. While the 1990s saw a significant expansion in U.S. innovation capacity in semiconductors, the nation’s production capacity continued to decline. In some cases, this owed to foreign countries out-competing the U.S. on labor costs. But more can be attributed to the significant subsidies foreign governments have been providing to build and maintain fabrication plans—a level of support that the U.S. hasn’t matched.  

As a result, companies such as Taiwan Semiconductor Manufacturing Company became global leaders in semiconductor production, while the U.S. largely lost its production edge. In 1990, the U.S. accounted for 37% of global semiconductor manufacturing capacity; by 2020, that number had fallen to 12%. 

Figure 1The U.S. semiconductor industry has not only lost its production leadership, but also the “learn by doing” processes that go with manufacturing. Numerous experts—ranging from Gary Pisano and Willy Shih to James Bessen to Alex Williams and Hassan Khan of Employ America—have shown that such learning allows constant product and process improvement, elevating competitiveness. This expertise is particularly important for firms on the cutting edge of product development or that are designing chips for emerging technology products such as autonomous vehicles. Generating such advances requires not just substantial iteration between designers and manufacturers, but often also new manufacturing technologies to develop and produce new chips at scale. At the same time, emerging technology products are increasingly employing customized chip designs that are best leveraged by collocating design and production.   

For a country like the U.S., which is consistently on the frontier of possibilities in production, such learning loss can undercut the nation’s innovation strength by both locking it out of the next generation of semiconductor technologies and compromising its ability to be a global leader in emerging technologies in other domains. This stagnation is can be seen in the sideways track of U.S. semiconductor industry employment (see Figure 2). 

Figure 2All of which is why the excitement surrounding Intel’s announcement in Ohio—while hugely welcome—must not divert attention from the broader and deeper campaign needed to rebuild U.S. semiconductor preeminence. As the current chip shortage revealed, decades of inattention have now created a crisis, showing the urgent need for substantial investment not just in scaling up the nation’s fragile production capacity but reinvigorating semiconductor research and development. 

The first steps in this project can begin immediately: The House of Representatives should move to pass the months-delayed America COMPETES Act (its rebrand of the Senate’s U.S. Innovation and Competition Act), and send it on to the House-Senate conference needed to pass it.   

Central to that act is the CHIPS for America Fund, which would provide over $50 billion to support domestic semiconductor manufacturing. Just under 80% of that funding ($39 billion) would go toward incentive programs to defray the costs of constructing facilities focused on semiconductor fabrication, assembly, and testing. A little over 20% of the fund ($11 billion) would go to funding research and development programs for advanced microelectronics. This would not only support research on the next generation of semiconductor design, but would also support breakthroughs in semiconductor manufacturing. 

CHIPS for America funding would also go toward establishing a National Semiconductor Technology Center for research on semiconductor design and manufacturing, as well as supporting startups and workforce development programs in the semiconductor space. It would also support programs aimed at enhancing domestic capabilities in advanced packaging (the process of using semiconductor assembly to drive further performance improvements), and provide funding for microelectronics research at the National Institute of Standards and Technology. In addition, the fund would support the creation of a new Manufacturing USA institute focused on semiconductor manufacturing.  

Other programs in the House’s America COMPETES Act and the Senate’s U.S. Innovation and Competition Act will also help in rebuilding the nation’s semiconductor competitiveness. A new Directorate of Technology and Innovation at the National Science Foundation could accelerate research and development and promote STEM workforce development. And a set of regional technology and innovation hubs could drive the emergence of deeper advanced technology ecosystems in promising regions—potent sites for new semiconductor production.  

While the CHIPS for America Fund and broader innovation legislation are a critical start, the bills by themselves won’t transform America’s semiconductor production competitiveness or alleviate supply chain challenges. The country needs to go further. 

For one thing, the America COMPETES Act does little to distinguish between investments that were already planned and those that may not have happened absent government funding. In this regard, the act’s provisions will need to be carefully implemented to induce new plant development and avoid becoming a windfall to companies for investments they were already planning to make. Likewise, as the United States International Trade Commission has noted, most back-end semiconductor assembly, testing, and packaging is outsourced to low-wage countries. If the U.S. wants to ensure a truly domestic supply chain, it will be necessary to invest in those capabilities as well. 

Additionally, while the CHIPS for America Fund’s investments are significant compared to the nation’s relative underinvestment over the past three decades, they still significantly lag the investments that other nations, particularly China, are making into this crucial technology. ITIF found that China supports its semiconductor industry to the tune of hundreds of billions of dollars in the form of subsidies, direct government ownership, preferential tax policies, and below-market financing, as well as through more nefarious means such as intellectual property theft, forced joint ventures, and forced technology transfer. It will be difficult for the U.S. to keep up with such tactics without even more significant investments. Steady, ongoing investment support will be necessary for years—not to mention trade policies aimed at ensuring that global competition in this industry plays out on private-enterprise-led, rules-governed terms, as World Trade Organization member nations have committed to.  

Moving forward, policymakers should consider a steady stream of ideas that would move the U.S. more explicitly toward what Williams and Khan call a true “industrial policy”—one that stresses both innovation and production. Such a strategy would pursue strong, government-led investments that create a diversity of good jobs in multiple manufacturing regions. As part of that, policymakers could enact a new generation of government purchasing agreements that guarantee a minimum demand of domestically produced advanced chips at any given time. Policymakers may also want to address stock buybacks and other forms of short-term financial engineering that have become commonplace across many industries, including semiconductors, and which undercut research and development investment and bias decisionmaking toward short-term returns.  

In short, if the U.S. wants to establish a domestic supply of semiconductors that both supports national security and improves overall supply chain resilience, policymakers will need to begin to move away from the short-term thinking that got us into the current situation.  

Intel’s Ohio announcement is a noteworthy development, but it does not by itself forecast a return to U.S. preeminence in semiconductor manufacturing. What would achieve that would be the hard work of building out a multiyear strategy to genuinely restore U.S. semiconductor production and learning at the technology frontier. Work like that can begin now with the CHIPS for America Fund and the America COMPETES Act, and should continue in the coming years as a sustained national priority.