China CHIPS Act Guardrails—The Next Great US Policy Innovation?
Will investment guardrails, or “money for national-security compliance,” be the next “export controls” or “secondary sanctions”?
The following article was written by Arrian Ebrahimi, Yenching Scholar and author of the Chip Capitols newsletter.
At what point does the ratio of cheese to pasta turn “macaroni and cheese” into “cheese and macaroni?” Thinking about the guardrails restricting expansion in China from the CHIPS and Science Act, I keep coming back to this tautology.
Is the CHIPS and Science Act a case of “grants with guardrails” or “guardrails with grants”? The CHIPS Act’s China guardrails are an innovative new policy tool, epitomizing the latest fusion of national-security and industrial policy. This tool allows Washington to bind domestic and foreign companies to the US’s national security agenda without organizing broad coalitions behind multilateral export control regimes or even unilateral outbound investment review. As CHIPS grants slowly start to flow after the March 31 rolling application review date, capitals around the world will watch American policy innovation in action.
After years of advocacy, semiconductor companies are undoubtedly thrilled by the CHIPS Notice of Funding Opportunity, which finally dropped on Tuesday. The $39 billion in federal funding for semiconductor manufacturing grants is certainly the largest part of the CHIPS Act, but the bill’s China guardrails also have the potential to revolutionize the US’s approach to economic competition.
Restoring American Semiconductor Dominance
Congress and the White House are blazing ahead with an unprecedentedly aggressive agenda to restore America’s strategic industries and limit China’s access to national security critical technology. To implement their plans, however, policymakers face a perennial dilemma: industry and allied governments are not nearly as keen to stop trading with America’s adversaries. During the Cold War, companies in Japan and Norway famously violated US export controls to sell advanced milling equipment to the Soviet Union, made possible in great part by Japan and Norway’s laxer export review policies. Today, American companies lobby heavily against limits on exporting to China, and allied governments haltingly follow the US’s China policy.
To sidestep the inevitable onslaught of lobbying and licenscing carveouts that trade restrictions typically spur, US lawmakers last year tried a different approach. The CHIPS and Science Act offers multibillion dollar grants for semiconductor manufacturing projects conditioned on guardrails prohibiting recipient companies from expanding manufacturing capacity in China for ten years. This incentive-based approach to encouraging compliance with US national security policy is deeply creative. [Jordan: if you’re the staffer who came up with this, reach out to ChinaTalk I’d love to connect off the record!] Companies usually lobby heavily against rules that limit their access to the Chinese market, and allied nations are often unwilling to sacrifice their domestic industries by cooperating with US export controls.
By offering grants in return for guardrails, the CHIPS and Science Act not only staunches chipmakers’ opposition to trade restrictions but also massages companies towards embracing the American policies toward China. Furthermore, by enticing foreign semiconductor companies to accept guardrails in return for grants, lawmakers spare US trade negotiators the arduous task of convincing allied governments to adopt the strictest parts of the US’s China policies.
Industrial Policy Born Out of Crisis
To understand Congress’ intentions with the CHIPS guardrails, it helps to recall the context in which the subsidy program was conceived: an immediate supply-chain crisis combined with longer-term concerns about China’s rise. This inspired industrial policy on a scale not seen since the space race.
At the onset of the COVID-19 pandemic, automobile manufacturers around the world canceled their orders for new chips, fearing a sudden drop in demand for new cars. After a few months, demand for cars rapidly recovered, but auto chip production lines needed months more to switch back into gear. Beyond demand shocks, a slew of natural disasters exacerbated the already tight supply for all types of chips. The February 2021 winter storm in Texas knocked out power for facilities across the state, including fabs run by Texas Instruments and Samsung. A March 2021 fire at a Renesas fab in Japan further strained automotive chip supplies, and TSMC had to cut production due to Taiwan’s spring 2021 drought.
These immediate supply chain crunches played into longer-term concerns over the US’s waning share of global semiconductor manufacturing and China’s rise. The US saw its share of global chip manufacturing fall from 37% in 1990 to 12% in 2020. Meanwhile, China’s share rose from 3% in 2000 to 15% in 2020.
While the PRC’s chip industry has long been limited to the least advanced ends of the supply chain, Washington took a collective gasp when the Chinese Semiconductor Manufacturing International Corporation (SMIC) announced it produced 7nm node chips for the first time. This placed SMIC just one generation behind the most advanced chips currently in commercial production globally; in absolute terms this accomplishment allows China to produce the high-performance chips required for servers and smartphones. While US restrictions on advanced manufacturing equipment will limit how well the PRC can implement 7nm production in the mid-term, Washington now faces a triple threat: China’s chip industry is catching up in 1) relative manufacturing capacity, 2) relative technological advancement, and 3) absolute technological advancement.
The Biden administration is responding to China’s rise with what it calls the “protect and promote” agenda. “Promote” means offering support to American industry (as with the CHIPS grants), and “protect” means cutting off adversarial nations’ access to US technology. While industry and even allied nations are thrilled by the idea of federal funds to grow the West’s chip industry, getting foreign companies and governments on board with the “protect” agenda is much more challenging.
Multi-Sectoral Restrictions Fail…
Lawmakers’ most recent crack at the “protect” agenda was a bipartisan bill establishing an outbound investment screening mechanism aimed at blocking any US investment in China that threatens national-security interests. The bipartisan National Critical Capabilities Defense Act (NCCDA) would have covered industries ranging from semiconductors and pharmaceuticals to the vaguely defined “artificial intelligence.”
Due to its broad scope, the bill faced adamant opposition from industry groups ranging from medical technology to software. Industry representatives negotiated the bill down from one that empowered the US government to cancel proposed foreign investments (as represented in the bill’s House version) to a softer one requiring only the reporting of certain foreign investments (as represented in the bill’s Senate version). Despite these concessions, the bill’s champions could not include it in the CHIPS and Science Act for fear of sinking the whole bill, and it remains to be seen if the 118th Congress will have any more success with such a broad effort.
…But a Deal with One Industry Succeeded
For externally observable strategic reasons, semiconductor companies were much more controlled in their critique of the NCCDA. As the broader business community fought outbound investment review legislation, policymakers and chipmakers danced a more intricate tango over CHIPS guardrails.
Cash is fungible. Fearing that CHIPS grants would simply offset the cost of further expansion in the PRC, China hawks in Congress insisted that recipients of federal funds be subject to certain guardrails. Most significantly, Section 103 of the CHIPS and Science Act restricts grant recipients for ten years from “the material expansion of semiconductor manufacturing capacity in the People’s Republic of China or any other foreign country of concern.” The bill permits the expansion of “legacy semiconductor” capacity primarily serving the Chinese market, which would leave China dependent on chips from US companies.
“Mandatory agreements” between the Department of Commerce and individual grant recipients will determine how the guardrails apply in any given case. In its February 28 Notice of Funding Opportunity, Commerce noted it would soon share additional details on the guardrails. This next announcement may lay out officials’ calculus for negotiating expansion restrictions with prospective grantees. For now, this is all the color we have:
Restricting trade for national-security reasons is usually a zero-sum game that spurs unmitigated opposition by all industries involved. By targeting restrictions to only companies receiving grants, however, lawmakers turned this dynamic on its head: companies can decide for themselves whether the economics make sense to sacrifice expansion in China in return for subsidized expansion in the US.
To that end, no semiconductor manufacturers — who would be the most natural opponents to the guardrails — openly rejected the imposition of China guardrails. Intel, for example, sought only to ensure the definition of “legacy semiconductor” remained flexible so older generations of technology could be progressively built in China as the leading edge advances. Since the CHIPS Act’s passage, the Semiconductor Industry Association, Intel, Micron, and Texas Instruments have all urged that the restrictions on expansion in China be limited to the wafer fabrication stage of manufacturing and exclude other stages like assembly, test, and packaging (ATP) — but they have not openly sought to undermine the underlying goal of the guardrails. As far as the rest of the industry was concerned, the guardrails are a non-issue. Fabless companies offered no pushback, as they do not operate physical manufacturing facilities subject to the restrictions; companies like Skywater, which manufacture chips only in the US, have even advocated for stronger restrictions that capture ATP as well as wafer fabrication.
Lastly, once chipmakers agree to guardrails in return for CHIPS grants, the marginal cost of complying with further US restrictions on business with China goes down. After the US in October 2022 rolled out a wide-ranging set of export controls hamstringing China’s access to semiconductor equipment and services, Intel CEO Pat Gelsinger told The Wall Street Journal that these restrictions were “inevitable” and “that’s why the rebalancing of supply chains is so critical.” Having already determined to apply for CHIPS grants and accept the guardrails, Intel’s business strategy now aligns much more closely with the US’s China policy and has relatively little to lose from further restrictions.
Sidestepping Allied Governments Via Allied Companies
The US’s most powerful tool for restricting China’s access to critical technology lies in export controls. For export controls to be effective, however, all countries with companies that provide the strategic technology in question must agree to restrict sales to China; otherwise, US suppliers will simply be replaced by foreign competitors. This process can take a long time, and, as reporting on decisions by the Dutch and Japanese governments indicates, the restrictions allies ultimately impose may not be nearly as stringent. [At the time of publication, news outlets have widely reported on an agreement by the Dutch and Japanese governments to join the US’s effort to restrict semiconductor equipment and talent exports to China, but the actual contents of this agreement have not been divulged.] The CHIPS and Science Act, however, once again sidesteps skeptics by paying for companies’ cooperation via grants.
By accepting CHIPS grants, foreign companies agree to tighter restrictions on business with China than any of their home governments require. As with the US manufacturers, foreign industries have advocated narrowing the guardrails. The Korea Semiconductor Industry Association (KSIA) urges that guardrail restrictions “not unduly impede” grant recipients’ existing facilities in China, but it recognizes “that the CHIPS Act requires effective limits on manufacturing in foreign countries of concern.” Similarly, TSMC supported the CHIPS and Science Act along with its guardrails, only urging that “expanding semiconductor manufacturing capacity at [its] Nanjing fab… be appropriately considered in time.” By putting foreign industries into the position of asking for exceptions to guardrails, the CHIPS and Science Act places US national security policymakers in a much stronger bargaining position than when they were just trying to convince foreign governments to impose restrictions.
An American Policy Innovation
No other US-allied, semiconductor-producing nation has implemented such a “money for national-security compliance” program in their chip subsidization scheme. This American exceptionalism primarily derives from the fact that no other semiconductor-producing region is as invested in limiting China’s advancement.
The article continues by exploring the semiconductor subsidies plans for Japan, South Korea, the EU, and Taiwan, exploring what strings those governments are planning to attach to money earmarked for industry.