U.S. Targeted Semiconductor Tariffs and Production-Linked Exemptions: Implications and Policy Responses

Executive Summary

U.S. Secretary of Commerce Howard Lutnick has formalized a policy of targeted tariffs on semiconductors coupled with exemptions linked to domestic production. While the tariff rates and implementation timeline remain undecided, the principle of using the presence of U.S.-based production facilities as the criterion for exemptions is already pressuring the industry to invest. Concurrently, the White House has designated Gyeonggi Province’s semiconductor belt as a high-risk area for transshipment, and cases of country-of-origin violations are being investigated in Taiwan, signaling a phase of simultaneous intensification of tariffs, export controls, and origin verification. The most likely scenario involves a differential tariff rate applied in proportion to the operational performance of Samsung Electronics’ Taylor fab and SK Hynix’s Indiana line. In this case, accelerating investment timelines and strengthening capabilities for origin verification will be key to determining the tariff burden. The South Korean government and companies must engage with working-level officials at the U.S. Department of Commerce from the earliest stages of designing the exemption criteria to establish existing investment records as a formal baseline. They also need to promptly establish a joint response channel to manage the three interconnected tracks of tariffs, verification, and export controls.

I. Situational Analysis

Situational Analysis: U.S. Targeted Semiconductor Tariffs and the 'Production-Linked Exemption' Policy

Background and Developments

On September 2, U.S. Secretary of Commerce Howard Lutnick announced in Washington that the administration is preparing a “targeted and thoughtful” tariff policy for semiconductors [1]. He reaffirmed the principle that tariffs would not be imposed on companies that manufacture products within the United States [1]. This statement is an extension of the trade policy stance that a potential second Trump administration has maintained since last year. The Office of the U.S. Trade Representative (USTR) has been touring manufacturing sites in states like Iowa, promoting “job recovery through reshoring” as a policy achievement [6]. Secretary of the Treasury Scott Bessent has also recommended the U.S. approach to tariffs to G20 nations, part of a recurring discourse within the administration that justifies tariffs as a tool of industrial policy [8].

The view of semiconductors as national security infrastructure dates back to the Biden administration’s CHIPS Act. At the time, the United States set a goal to reverse the decline of its share in global chip production from 40% in the 1990s to around 10% [3]. The Trump administration is inheriting this framework but replacing subsidies with the coercive tool of tariffs. The repeated mention of TSMC and Micron, which have made large-scale investments in Arizona, in Lutnick’s remarks shows that they are being used as model cases of compliance with onshoring pressure.

Current Situation

Lutnick’s statement remains at the stage of outlining a general direction, as it did not specify concrete tariff rates or an implementation timeline [1]. However, the principle of “differential application for companies without U.S. production facilities” is already functioning as pressure on the entire semiconductor industry to expand investment. Around the same time, the U.S. Department of Commerce was also confirmed to be reviewing measures to restrict China’s access to advanced computing via overseas data centers [5][17]. This is a move to expand the scope of controls beyond the physical export of semiconductors to include remote cloud access. This increases the compliance burden for cloud providers and data center operators in Southeast Asia and suggests a phase where the twin pillars of tariffs and export controls are being simultaneously strengthened.

In Taiwan, separate from the tightening of tariffs and controls, suspicions of country-of-origin laundering are coming to the fore. It was reported by local Taiwanese media that IC substrate manufacturer Unimicron was investigated by prosecutors on suspicion of origin violations [15]. In response, the chairman of Gudeng Precision stated that “trust is the core foundation of Taiwan’s position in the supply chain,” attempting to quell industry concerns [16]. This demonstrates that U.S. onshoring pressure is producing the side effect of intensifying origin verification across the entire supply chain, going beyond simply attracting investment. South Korea is no exception to this trend. The White House Office of Trade and Manufacturing Policy’s August 13 report, “The Great Transshipment Scam,” specifically identified South Korea’s Gyeonggi semiconductor belt as a high-risk region with deep integration into supply chains serving China [7].

Key Actors and Positions

U.S. Department of Commerce (Lutnick)is designing semiconductor tariffs as an incentive for domestic production. Lutnick’s pattern of remarks suggests an approach that uses tariffs not as a punishment but as negotiating leverage to influence investment decisions [1]. The investments in Arizona by TSMC and Micron are repeatedly cited as evidence that this approach has already yielded some success.

USTR (Ambassador Greer)is framing tariff policy as a domestic political achievement that leads to wage increases and job reshoring [6]. It is also pursuing measures to enhance price transparency for critical minerals like rare earths, concurrently implementing trade and industrial policies aimed at reducing supply chain dependence on China [12].

U.S. Department of the Treasury (Bessent)is presenting tariffs not as an exceptional U.S. measure but as a universal defensive tool that the entire G20 should adopt [8]. This suggests the possibility that the discourse on semiconductor tariffs could expand beyond a bilateral U.S.-China issue to a broader debate on reshaping the multilateral trade order.

Taiwanese Semiconductor Industry (TSMC, Unimicron, Gudeng Precision)shows a clear divergence between TSMC, which has already managed tariff risks through U.S. investment, and downstream industries (e.g., substrates) exposed to origin verification risks [15][16]. The Taiwanese government and industry are concerned that damage to supply chain credibility could weaken their negotiating position with the United States.

South Korean Government and Semiconductor Industryare in a passive position, anxiously monitoring the situation following Lutnick’s remarks. Although both Samsung Electronics and SK Hynix already have or are building production facilities in the United States, the specific scope of the “production link” that will serve as the basis for exemptions—such as the share of U.S. production in total output or the inclusion of cutting-edge processes—remains unclear.

Chinais an indirect target of this measure but has also been identified as a potential unintended beneficiary. The Peterson Institute for International Economics (PIIE) has analyzed that Trump’s next round of tariffs could paradoxically lead some companies to return to China [2]. Empirical analysis is also presented showing that past tariffs on China resulted in shifts to bypass routes rather than a genuine reduction in dependence on China [14].

Key Issues

The first issue is the baseline for tariff exemptions. It has not been determined whether the criterion will be the mere “presence” of production facilities in the U.S. or if it will also include production volume and process technology levels. How this standard is set could create disparities in benefits even among companies that have already invested in Arizona and Texas.

The second issue is the overlapping risk of tariffs, transshipment monitoring, and export controls operating simultaneously. Although the three policy tracks—targeted semiconductor tariffs, designation of high-risk countries for transshipment to China, and restrictions on remote computing access—are proceeding separately, they are in fact interlinked, creating a structure that cumulatively increases the response burden on South Korean companies [5][7].

The third issue is the policy’s unpredictability. Lutnick’s remarks only offered a general direction, lacking detailed regulations at the executive order level. The PIIE points out that this uncertainty itself is a factor that distorts corporate decisions on investment and production locations [2].

II. In-Depth Analysis

In-Depth Analysis: U.S. Targeted Semiconductor Tariffs and the 'Production-Linked Exemption' Policy

1. Root Cause: The Structural Vulnerability of the U.S. Semiconductor Manufacturing Base

The root of this measure lies in the decades-long decline of America’s semiconductor manufacturing capabilities. The U.S. share of global chip production has fallen from 40% in the 1990s to about 10% today [3]. The consensus in Washington policy circles is that this gap cannot be closed by subsidies alone. The Biden administration attempted to restore manufacturing competitiveness by injecting approximately 50 trillion won through the CHIPS Act [3]. However, subsidies are a slow instrument, requiring time for congressional budget deliberations and corporate investment decisions. The Trump administration seeks to solve this speed problem with the immediate coercive tool of tariffs. Secretary Lutnick’s repeated emphasis on the principle that “companies that produce in the U.S. will not pay tariffs” signifies the design of an incentive structure based on tariff avoidance rather than subsidies [1]. From a corporate perspective, this creates a structure where the freedom not to invest is effectively offset by the cost of tariffs.

The fundamental motivation for this approach is a shift in Washington’s perception, which now defines semiconductors not as an object of industrial policy but as a national security asset. If semiconductors were once the “rice of industry,” they have now been elevated to the status of “infrastructure” [3]. This definition—as the foundation of the national economy and the bedrock of security—provides the rationale for applying a security logic, rather than the conventional logic of comparative advantage, to tariff policy. It means that geopolitical risk takes precedence over economic efficiency in calculating tariff rates.

2. Structural Context

Political Structure. The Trump administration’s tariff policy is intertwined with a political narrative of restoring domestic manufacturing jobs. The USTR’s tour of a tire factory in Iowa to promote “wage increases and job reshoring” illustrates that this policy functions as a domestic political asset, not just a pure trade strategy [6]. Treasury Secretary Bessent’s recommendation to G20 countries to adopt the U.S. approach to tariffs is part of the same context [8]. A discourse justifying tariffs not as an exceptional U.S. tool but as a universal instrument of industrial policy is being repeated within the administration. This political pressure constrains flexibility in the detailed design of the semiconductor tariffs. For an administration that needs to continuously present “onshoring achievements” during a midterm election cycle, granting exceptions to some companies risks undermining the entire policy’s persuasiveness.

Economic Structure. The semiconductor supply chain is a multi-stage system of division of labor, spanning design, foundry, packaging, and materials. Relocating a specific process to the U.S. does not automatically reconfigure the entire supply chain. As a Microsoft Azure hardware executive pointed out, there is a view within the semiconductor industry that simply expanding memory production capacity will not resolve supply chain bottlenecks for AI infrastructure [13]. This means that even if tariffs succeed in inducing the relocation of production facilities, additional time and investment are needed to actually alleviate supply shortages. PIIE analysis concludes that tariffs on China, despite being in place for years, failed to structurally reduce U.S. dependence on Chinese suppliers [14]. This was because de facto dependence continued through bypass routes. This precedent suggests that semiconductor tariffs may also succeed in inducing a superficial shift in production locations but are likely to have a limited effect on the substantive reorganization of the entire supply chain.

Security Structure. Separate from tariffs, the U.S. Department of Commerce is reviewing restrictions on China’s remote access to computing power via overseas data centers [5][17]. This is a move to expand the scope of control beyond the physical export of chips to include cloud-based access. It shows that semiconductor policy is evolving from simple tariffs into a security control network that encompasses all stages of production, distribution, and use. The White House Office of Trade and Manufacturing Policy’s “Great Transshipment Scam” report, which designated the Gyeonggi semiconductor belt as a high-risk area, is part of this same structure [7]. Production site verification and country-of-origin tracing are becoming essential prerequisites for tariff policy.

3. Historical Precedents and Similar Cases

The most direct precedent is the tariff policy toward China during the first Trump administration. The PIIE analyzes that these tariffs, instead of actually reducing bilateral U.S.-China trade, resulted in an increase in bypass exports via third countries [14][2]. This means that the route of Chinese-made components being assembled in third countries and then re-exported to the U.S. expanded. The current semiconductor tariffs could create similar incentives for circumvention. The PIIE warns that the recent combination of expanded tariffs on 16 trading partners and relaxed tariffs on “non-sensitive” Chinese goods could paradoxically cause companies to return to China [2]. If the production-linked exemption creates a structure that encourages bypass production in certain countries, the intended onshoring effect of the policy could be diminished.

The case of suspected origin laundering in Taiwan is another relevant precedent. The investigation of IC substrate manufacturer Unimicron by prosecutors for alleged origin violations reaffirms the historical pattern that as U.S. pressure intensifies, so do the incentives for bypass production and origin laundering [15]. The attempt by the chairman of Gudeng Precision to calm industry fears by stating that “trust is the key to Taiwan’s position in the supply chain” reflects concerns that such pressure could shake the foundation of trust for the entire industry [16].

A comparison with the CHIPS Act is also instructive. The Biden administration’s subsidy-based approach was predicated on voluntary corporate investment decisions. In contrast, the Trump administration’s tariff-linked approach is a coercive structure that explicitly imposes a cost for not investing. While both policies aim for the same goal of “production in the U.S.,” the nature of the instrument has shifted from incentive to coercion. The repeated mention of TSMC and Micron in Lutnick’s remarks, after they have already committed to their Arizona investments, shows that this shift is functioning as a reward for companies that have already complied and a warning signal for those that have not yet decided.

4. Key Variables Shaping Future Developments

The first variable is the timing for the specification of tariff rates and application criteria. Lutnick’s remarks remain at the level of principle, with no specific rates or executive orders yet released [1]. This interim period itself acts as a cost of uncertainty for companies.

The second variable is the criteria for determining the “presence of production facilities.” The response strategies of Samsung Electronics and SK Hynix will differ completely depending on whether simple assembly and packaging facilities qualify for exemptions, or if substantial production at the wafer fab level is required.

The third variable is the degree of linkage between strengthened origin verification and tariff policy. If discussions on designating high-risk countries for transshipment are combined with the implementation of semiconductor tariffs, the customs clearance process in the U.S. for products from the Gyeonggi semiconductor belt could itself be delayed or require additional documentation [7].

The fourth variable is the interaction with separate control networks related to China. If discussions on restricting remote data center access materialize [5][17], even if South Korean companies secure U.S. production facilities, business segments with a high share of revenue from China will face separate regulatory risks. Receiving tariff exemptions and being excluded from export controls are likely to proceed on separate tracks.