The CHIPS Program Office Needs to Think Like Economic Developers, Not Bankers
In passing the CHIPS Act of 2022, Congress signaled its determination to revitalize U.S. semiconductor competitiveness by authorizing a 25 percent investment tax credit and appropriating $52.7 billion to support the industry. That included $11 billion for research and development (R&D) activities and $39 billion for a “CHIPS for America Fund” to provide financial incentives to build, expand, and equip domestic fabrication facilities.
This was sorely needed, because the U.S. share of global semiconductor production had dropped from 37 percent in 1990 to 12 percent in 2021 while other nations were expanding their incentive packages to attract fabs. Yet, despite the urgency of restoring U.S. semiconductor manufacturing competitiveness, 16 more months have now passed, and the CHIPS Program Office so far has disbursed only a small fraction of the funds Congress authorized. The program’s administrators need to accelerate the process by thinking more like economic developers—whose MO is “let’s get it done”—and less like bankers who dot every i and cross every t hoping to avoid risk and limit liability.
The Legislation Certainly Has Been Popular
Dozens of semiconductor manufacturers have announced U.S. investments in the wake of the CHIPS Act. Among others, this list includes an Intel commitment to invest at least $20 billion in two new Ohio semiconductor fabs; TSMC’s announcement that it plans to triple investment in its Arizona fab to $40 billion and build a second, 3 nanometer (nm) semiconductor fab; and Micron’s announcement that it would build a megafab for memory chips near Syracuse, New York that may entail as much as $100 billion in investments over the next 20 years.
In total, the U.S. Department of Commerce has received more than 570 statements of interest and more than 170 pre-applications, full applications, and concept plans for CHIPS incentive programs. And by January 2024, companies had announced over $231 billion in commitments in semiconductor and electronics investments in the United States. Many of these companies quickly began putting shovels in the ground toward starting construction of fabs with the expectation that they’d be in line for a share of incentive funding from the legislation.
The semiconductor investment incentives can be helpful in arresting the loss of U.S. semiconductor manufacturing because they help to defray or to match the nature of other countries’ investment attraction incentives such as those that defray the cost of land, energy, construction, or transportation links. Such foreign government incentives (in countries such as Korea or Taiwan) can reduce up-front capital expenditures for land, construction, and equipment and may offset 15 to 40 percent of the gross total cost of ownership (TCO) of a new fab (pre-incentives), depending on the country. Indeed, this was one of the key justifications of the CHIPS legislation, because, as the Boston Consulting Group (BCG) found, the 10-year TCO of U.S.-based semiconductor fabs is 25 to 50 percent higher than in other locations, with government incentives in those countries accounting for 40 to 70 percent of the U.S. TCO gap.
The semiconductor investment incentives can be helpful in arresting the loss of U.S. semiconductor manufacturing because they help to defray or to match the nature of other countries’ investment attraction incentives.
The first two awards that the CHIPS Program Office has announced (pending completion of a due diligence review) went to BAE Systems—$35 million to modernize its Microelectronics Center, which makes chips for defense needs, such as F-35 fighters—and to Microchip Technology—$162 million to support the company’s production of microcontroller units for aerospace, automotive, commercial, industrial, and defense customers. Also, on January 31, 2024, the CHIPS Program Office filed a Notice of Intent initiating competitions to create a National Advanced Packaging Manufacturing Program and a CHIPS Manufacturing USA Institute.
While this certainly represents progress, the CHIPS Program Office needs to significantly accelerate its output, focusing more on the big picture and not the minutiae of deal terms. In other words, it needs to start acting more like an economic developer than a banker. That’s in part because semiconductor fabs are extremely significant investments for semiconductor manufacturers, for whom the time value of money is extremely important, meaning they need to get new fabs into operation as quickly as possible to begin recouping their investment.
For the reality is that semiconductor manufacturers encounter significant upfront capital costs in even breaking ground on a new fab, let alone completing its construction. In fact, the costs to construct a leading-edge 3 nm fab are now approaching $20 billion. Moreover, the cost to build a new fab gets more expensive every day. For instance, a recent BCG report found that, “a fab completed in 2026 would carry a ten-year TCO of $35 billion to $43 billion—33% to 66% higher than today’s costs.” The BCG report reached two conclusions:
First, the cost of building and operating these fabs has risen significantly. Second, while the government incentives landscape is still evolving across multiple parts of the world, new, more aggressive financial support programs will be critical to make the economics of these massive capital-intense projects viable.
The point is that if the U.S. government is serious about rapidly stimulating U.S. semiconductor manufacturing production, then it needs to start disbursing funding as soon as shovels hit the ground and to do it with few strings and red tape attached. This has in fact been more indicative of the approach coming out of Europe. One company receiving semiconductor incentives from a European government reported the timing of the disbursements were roughly one-third upfront, one-third the following year, and the final one-third a year later.
Too Much Red Tape, Not Enough Shoveling
In contrast, the CHIPS Program Office appears to be making the timing of the disbursement of funds contingent on applicants meeting scheduled milestones. In protecting taxpayer interests and ensuring their money is well spent, the U.S. government is certainly justified in making investments judiciously and ensuring that recipients achieve the milestones and deliverables they’re committing to. But this needs to be in broad strokes and not excruciating detail that entails significant compliance costs. Moreover, it needs to recognize that the grants, loans, or loan guarantees chip makers receive will represent only a small fraction of the total capital expenditures they’ll be making as they build or expand their operations. In fact, the CHIPS Program Office’s Notice of Funding Opportunity (NOFO) notes that, “the Department generally expects most CHIPS Direct Funding awards will range between 5% to 15% of project capital expenditures.” In other words, semiconductor manufacturers aren’t going to be making these investments lightly—they’re putting a significant amount of their capital at risk as well—which should allay to some degree taxpayers’ (and thus CHIPS Program Office) concerns handing out cash only to see the facility not ending up producing chips. If the Department of Commerce were a bank making a loan, perhaps this level of green-eye-shade accounting might make sense. But the United States is working to convince companies to make investment choices in the United States when they have many more choices overseas, many of them with support from economic-development-type governments, not banker types.
Further, it does appear that disbursement of CHIPS incentives is being bogged down in the arcane minutiae of deal terms. As The Wall Street Journal reported on January 18, 2024, “Negotiations between the Biden administration and TSMC over subsidies have proven challenging, say people on both sides. Issues include both the amount of subsidies and the conditions to which TSMC must agree to receive the money, they say.” Similarly, a Business Korea article noted in December that Samsung recently decided to delay its $17 billion investment in a chip fab outside Austin, Texas with the delay “speculated to be due to issues related to US government subsidies and various permit complications.”
In protecting taxpayer interests, the U.S. government is certainly justified in making investments judiciously and ensuring that recipients achieve the milestones and deliverables they’re committing to. But this needs to be in broad strokes and not excruciating detail.
One particularly concerning element of CHIPS incentive packages that has been referenced in several news reports pertains to “upside sharing.” CHIPS incentives applicants must provide the Department of Commerce with projected cash flows over the lifetime of their project. As the CHIPS NOFO states, “Recipients receiving more than $150 million in CHIPS Direct Funding will be required to share with the U.S. government a portion of any cash flows or returns that exceed the applicant’s projections (above an agreed-upon threshold specified in the award).” That the government might be able to go back (even potentially years later) and contend that a semiconductor company’s investments ended up being too successful and now the company must return a share of revenues to the government would certainly have a detrimental impact on the effective use of the incentives instrument. For instance, accurately projecting future cashflows with the risk of government taking profits back is likely a very difficult thing for companies to do, and adhering to a policy like this may simply be unmanageable for potential grant recipients.
It appears that this upside sharing requirement was added by the Department of Commerce, presumably under Section 9909, “Additional Authorities,” of the CHIPS Act, which gives the Secretary of Commerce the ability to “enter into agreements, including contracts, grants and cooperative agreements, and other transactions as may be necessary on such terms as the Secretary considers appropriate” (i.e., this author could not discern the upside sharing requirement in the legislative text). This kind of provisions appears to reflect a mindset that these companies need to be in the United States and that the Department of Commerce is doing them a favor by providing grant awards. The reality is almost the opposite.
In addition, the CHIPS Program Office is attaching a detailed list of milestones and covenants to the disbursement of CHIPS for America Fund incentives, including a number that relate far more to social policy objectives than stimulating U.S. semiconductor manufacturing competitiveness. This includes providing information in their grant applications pertaining to issues such as: validating that the applicant has developed “a plan for access to child care for facility and construction workers”; promoting “community investments” including “financing or building affordable housing or providing housing vouchers”; and ensuring their use of “domestic manufacturing and content” where applicants describe “whether and how they plan to utilize iron, steel, and construction materials produced in the United States” as part of their projects. Unfortunately, such a grab bag of social policy objectives—which Ezra Klein has called the “everything bagel” of progressive red tape—many of them inserted by congressional legislators, has clearly had an impact in slowing the disbursement of CHIPS Act funding and again reflects the notion that the U.S. government is helping these companies, rather than the opposite reality of the fact that they are helping the United States.
Another good example of the overengineering of CHIPS Act funding requirement (although responsibility for this one lands at Congressional feet) is that it requires companies to comply with Davis-Bacon Act stipulations in most contracts. The Davis-Bacon Act sets wage levels that apply to contractors and subcontractors performing on federally funded or assisted contracts in excess of $2,000 for the construction, alteration, or repair of public buildings or public works. Certifying to the government that the payrolls of contractors and subcontractors involved in building fabs with CHIPS incentives are Davis Bacon Act-compliant adds a whole new layer of bureaucratic complexity and expense. It also raises the challenge that many companies moved immediately in the wake of CHIPS Act passage to start constructing fabs; but now if, to receive CHIPS funds, they must go back and renegotiate contracts already agreed to in order to be in compliance with Davis-Bacon requirements, this only further adds delays.
Another concern pertains to the environmental review process that will apply to semiconductor facilities built with CHIPS Act funds. CHIPS Act applicants were required to submit an “A-Z” environmental questionnaire to the Department of Commerce. But as The Wall Street Journal reports:
The most immediate threat to the timely construction is the National Environmental Policy Act [NEPA], which requires large federally funded projects to pass environmental review before grants are released, regardless of whether they have already obtained state and local government permits. Full NEPA reviews took an average of 4.5 years between 2013 and 2018, according to a federal government report.
Observers contend that each year of delay in receiving environmental permitting review adds roughly 5 percent to the cost of a constructing a semiconductor chip plant. Commerce Secretary Gina Raimondo herself has warned that U.S. efforts to build out the domestic semiconductor industry could be delayed by years if companies are required to go through standard environmental reviews. As she commented in a recent Bloomberg News interview, “Obviously we want to do everything always to protect the environment. … But this is a national security priority, and we need to move quickly.“
Thomas Hochman, in an article titled “It’s Not Just NEPA: Reforming Environmental Permitting,” offers a useful policy proposal here that should be considered:
When permitting semiconductor facilities, authorities at the state and local levels should treat the entire fab as the regulated emissions source, rather than treating each piece of fab equipment as a unique emitter. Right now, the latter approach is more common, but it’s enormously prohibitive. Semiconductor facilities must constantly swap out and replace equipment to keep up with operations. When each piece of fab equipment is treated as its own source of emissions, facilities must notify permitting authorities prior to every single change they make.
The Competition Isn’t Letting Up
Certainly, there are many thorny challenges to tackle, but they need resolution soon, for the clock is ticking, and other countries are moving rapidly with their investments, which have enabled chip manufacturers to stand up plants quicker elsewhere. For instance, a TSMC facility in Japan, which was announced after the company’s breaking ground on its Arizona facility, is nearing completion, even as completion of TSMC’s Arizona fab has been pushed back to 2025. It’s not as if these companies are building chip factories in Yosemite National Park. In virtually every case, they are building factories in already built-up metropolitan areas. The idea that they have to comply with federal environmental reviews is frankly nonsensical. State and local governments are perfectly capable of protecting their local environment.
Countries are hungry, and they’re moving rapidly to attract globally mobile semiconductor industry investment.
Meanwhile, the global competition to attract semiconductor-sector investments only intensifies. Japan announced $21 billion in funding in 2023 to support its semiconductor industry, including a $3.5 billion subsidy for a new 10nm to 20nm fab in the Kumamoto prefecture. Taiwan’s Chips Act, passed in January 2023, offers investment tax credits of 25 percent on R&D and 5 percent on equipment. South Korea’s K-Belt strategy seeks to build the world's best semiconductor supply chain by 2030 with a $450 billion investment plan. It offers semiconductor manufacturers tax credits that cover up to 25 percent of facility costs and 30 to 40 percent of R&D expenditures. China invested $290.8 billion in semiconductor projects between 2021-2022 alone. In April 2023, Europe approved its $47 billion European Chips Act, which includes investment incentives, R&D funding, and talent building measures. Germany has allocated $20 billion for semiconductor incentives. India has made at least $10 billion available through its “India Semicon Programme” and is offering incentives of up to 70 percent (50 percent from the federal government, 20 percent from a relevant state) on an up-front (or “pari passu”) basis, making India’s incentives perhaps the world’s most generous at the moment. In short, countries are hungry, and they’re moving rapidly to attract globally mobile semiconductor industry investment.
As tech entrepreneur Marc Andreessen famously said in the midst of the coronavirus pandemic, “It’s Time to Build.” Accelerating the construction of semiconductor facilities here in the United States should be the paramount goal, so it’s time for policymakers to start thinking like economic developers with a “let’s get it done” mentality and not bankers with a “let’s cover every base and limit every liability”-type mindset. The Department of Commerce and the CHIPS Program Office are to be commended for the care, detail, and professionalism with which they have taken to their important task; but at the same time, the funding needs to be getting more quickly out the door with fewer impediments. Otherwise, we risk continuing to lose out to other nations in semiconductor production.