Confronting intensifying strategic competition with China, the United States is advancing its semiconductor manufacturing capabilities, aiming to restore technological leadership and mitigate national security risks, while reducing overdependence on fragile global supply chains.¹ Along with growing geopolitical fissures, global supply chain shocks during the COVID-19 pandemic revealed to Washington the extent of its vulnerabilities with regard to offshore chip production and jolted policymakers into a renewed push for domestic resilience.² Much good work has been undertaken (see the CHIPS Act), yet a critical bottleneck in chip production remains underemphasized: Assembly, Test, and Packaging (ATP).³ As the final stage of chip production, ATP plays a decisive role in determining circuit speed, heat dissipation, and reliability. Without packaging, chips themselves are useless.⁴

To strengthen American leadership in advanced packaging technologies, the National Advanced Packaging Manufacturing Program (NAPMP) was established under the CHIPS for America initiative. Allocated approximately $3 billion, NAPMP works to develop a self-sustaining, high-volume domestic advanced packaging industry, ensuring that advanced node chips are both manufactured and packaged within the United States.⁵

These investments underscore the federal commitment to enhancing ATP capabilities domestically — a priority that has seen continuity across administrations. (While the Trump Administration has criticized aspects of the CHIPS Act, it has also taken steps to accelerate ATP investment through mechanisms like the U.S. Investment Accelerator and support for facilities such as the TSMC-Amkor site in Arizona.)⁶ But several important questions remain: How well will this solve current supply chain stability issues? Will these investments wean the United States from Asian ATP dominance? Is the U.S. the only desirable location for ATP investment?

In short, the United States must decide what to onshore, and how.

Given the current state of American ATP capacity, the dominance of Asian players in this sector, and the strategic imperatives for onshoring or nearshoring ATP processes, a three-tiered onshoring strategy is required, one aligned with national priorities and firm-level incentives. Increased investment in ATP automation and a counter-intuitive, politically fraught, yet viable industrial collaboration with Mexico, would help the United States to address some of the more salient challenges.

The Current State of Assembly, Test, and Packaging: Critical Bottlenecks, Asian Dominance

The United States houses less than 3 percent of global ATP capacity by revenue.⁷ Putting this into perspective, American firms account for more than 50 percent of global chip design revenue, as well as 10 percent of global manufacturing capacity, which is projected to grow to 14 percent for all chips and 28 percent for advanced chips by 2032.⁸ This comparative disadvantage is fueled by limited ATP infrastructure, non-competitive cost structures, and a historical overreliance on Asian players. The CHIPS Act has spurred some domestic efforts, but ATP-dedicated funding comprises only roughly 4.81 percent of the total $33 billion in funding allocated so far.⁹

Asian countries lead in ATP due to high-volume, cost-effective production, creating the most attractive investment environment for firms. As of 2023, Asia accounted for 95 percent of the ATP facilities in the global market.¹⁰ In the past two years alone, this market has seen:

• In Malaysia: A total investment of $7 billion to expand current Intel operations in Penang, including the development of ATP cleanroom space.¹¹
• In Vietnam: $900 million of investment into legacy ATP through 2026 and $1.6 billion of investment into advanced packaging facilities.¹² Also, in Vietnam, a $1.6 billion Amkor investment, including the development of 600,000 square feet of new cleanroom space.¹³
• In India: $2.75 billion Micron investment, developing 500 thousand square feet of cleanroom space.¹⁴

While exact facility capacity is difficult to pin down, considering the volume of investment and the relative scope of cleanroom size, it would be virtually impossible for the United States to achieve capacity or cost parity with Asian ATP players.

The most promising investment into American ATP capacity so far has been Amkor in Peoria, Arizona, which was a direct result of the National Advanced Packaging Manufacturing Program, the ATP-dedicated branch of the CHIPS Act. Amkor is the second-largest Outsourced Semiconductor Assembly and Test (OSAT) firm in the world, with market share estimates ranging from 10-15 percent, led only by Taiwanese giant ASE.¹⁵ In 2024, it was announced that the U.S. government would allocate $400 million in direct subsidies as well as a 25% tax credit (through 2026) to support Amkor’s $2 billion investment into a 500,000 square foot cleanroom and facility.¹⁶

A major factor for Amkor’s decision to enter into the U.S. market was the creation of a three-way partnership between Amkor, Apple, and TSMC. TSMC’s advanced packaging technologies—Chip-on-Wafer-on-Substrate (CoWoS) and Integrated Fan-Out (InFO)—are in extremely high demand, leading to a critical bottleneck in semiconductor packaging.¹⁷ To address this, TSMC has entered a turnkey agreement with Amkor, allowing Amkor to use these proprietary processes for chips fabricated by TSMC in Arizona.¹⁸ This collaboration gained further momentum with Apple’s announcement that it will become Amkor’s newest and largest customer in Peoria, underscoring the strategic importance of domestic packaging capacity in the U.S. semiconductor supply chain.¹⁹

Impetus and Strategies for Nearshoring: Mexico

While the United States lacks the possibility of cost or capacity parity with Asia, its close neighbor in Mexico–primarily in the state of Jalisco–offers advantageous cost structures, strong semiconductor/automobile industry presence, and most importantly, growing interest from some of the world’s leading tech firms.

Mexico is an increasingly attractive nearshoring destination for semiconductor ATP operations, thanks to its existing OSAT facilities, strong chip consumption, close trade ties with the U.S., and a cost-efficient labor force.²⁰ The $100 billion in U.S. foreign direct investment and the integrated manufacturing ecosystem enabled by the USMCA further bolster its appeal.²¹ However, this potential is tempered by structural challenges: Mexico’s microelectronics ecosystem remains underdeveloped relative to Asian competitors like Vietnam or Malaysia, and many critical components still need to be imported from Asia.²² Its history of weak tech-oriented incentives—illustrated by the underwhelming 2023 semiconductor tax package—raises concerns about the government’s ability to support the sector effectively. Moreover, while labor is abundant, the shortage of high-skilled manufacturing talent and a thin training pipeline limit Mexico’s readiness to scale with the demands of a growing global semiconductor industry.²³

There is a strong and growing corporate presence in cutting-edge technologies developing in Mexico. Foxconn is building NVIDIA Blackwell servers, most likely in Guadalajara, Jalisco, a city that houses 70% of Mexico’s semiconductor industry.²⁴ The plant will, according to Foxconn’s representatives, be the largest GB2000–one of the world’s most advanced AI computing systems, delivering up to 30 times faster performance than previous generation-GPUs–production facility in the world.²⁵ ASE and Micron are also looking to build ATP and memory facilities, respectively, in Jalisco, likely to be completed within the next five years.²⁶

The U.S. government has also taken several steps to prioritize Mexico as a nearshoring destination. It has explicitly tapped the State Department’s “International Technology Security and Innovation Fund” to promote ATP development strategies in Mexico, either through workforce development programs or infrastructure development strategies.²⁷ USAID, in partnership with its Mexican counterpart, FUMEC, has identified the top five Mexican states best-suited for high-tech investment.²⁸ Finally, in September of 2024, the U.S. Department of State and Mexico’s Secretariat of Economy co-hosted a conference named the “Americas Partnership Semiconductor Symposium: Expanding the Supplier Ecosystem in Mexico City.”²⁹

It is prudent to consider the Trump Administration’s new tariff regime and the international outcry it has provoked. While President Claudia Sheinbaum has emphasized the importance of avoiding a “tit-for-tat” trade war—opting instead to focus on strengthening Mexico’s economy—the long-term implications of these tariffs for bilateral semiconductor cooperation remain unclear.³⁰ At the very least, we can expect the Mexican government to prioritize negotiating better terms and preserving the trade status quo, rather than pursuing new partnerships under unfavorable conditions.³¹ The administration’s broader posture toward trade and techno-industrial collaboration may further dishearten Mexican technology firms, dampening their previously positive outlook on collaboration with the United States. As a result, the viability of Biden-era efforts to build a tighter U.S.-Mexico semiconductor ecosystem is increasingly uncertain.

The Three “Tiers”

Given the difficulty of matching Asian investments into ATP, the United States should not look to provide a competitive investment landscape for ATP firms. Leading ATP companies are unlikely to engage extensively in U.S. development, as this would cause direct competition for facilities in Asia. The high U.S. industrial wage and low profit margins of ATP manufacturers also make it difficult to create a vibrant ATP industry in the U.S.³² Therefore, the United States must make hard decisions and decide  what to onshore, and how.

To prioritize effectively, the United States must focus its ATP onshoring strategy on areas where national security needs and firm-level profitability most clearly align. The following tiered framework—developed in consultation with former and current ATP executives—offers a targeted approach to identifying which segments of the chip ecosystem are most suitable for domestic packaging. It balances strategic imperatives with commercial viability, offering a practical path forward in a landscape where full ATP onshoring remains unrealistic.

First Tier: High-Performance Computing, Data Servers, AI Applications. Integral to national defense and data security, ATP executives felt these technologies to be the most beneficial for both the U.S. government and onshoring firms to prioritize. The production and packaging of cutting-edge technology to be used in high-performance computing or data servers could be the most important thing to effectively secure through domestic supply chains. The complexity of this technology provides profitable opportunities for ATP firms.

Second Tier: Automobiles. As noted by the Bureau of Industry and Security in a recent Notice of Proposed Rulemaking, the presence of foreign, potentially malicious legacy technologies in both critical infrastructure fleets (military, law enforcement, medical, energy) and consumer vehicles poses a cyber-attack risk. ATP firms could specifically support the onshoring of chips used in ADAS (Advanced-Driver Assistance Systems) and V2X (Vehicle-to-Everything Connectivity Systems) highlighted as risk-exposed by the BIS.

Third Tier: Cell Phones. The majority of cell phones on the market do not require more difficult (and therefore more lucrative) processes of manufacturing lines from ATP firms.³³ Here again, however, the Amkor-Apple-TSMC agreement comes into play, as it is likely that Apple’s most advanced mobile chips are complex enough to provide incentives for Amkor to package them.

The Peoria Plan: A Strategy for Deciding What to Onshore, and How

Given the structural limitations of the U.S. manufacturing environment, fully onshoring ATP is not a feasible goal. However, with carefully designed incentive structures and investments in workforce development, the United States can and should build domestic capacity for the most critical segments of ATP—particularly those tied to AI, defense, and high-performance computing. Achieving this will require a deliberate mix of public-private coordination, targeted subsidies, and ecosystem-building efforts, including:

• The coverage of overhead costs through direct government investment.
• A long-term plan for decreasing the need or reliance of companies on tax breaks, ideally through significant and diminishing tax credits that “wean” ATP firms off subsidies gradually, rather than removing them abruptly.
• Development of stronger public-private partnerships within the CHIPS Act Office. Through partnerships between the USG and private players, the most effective incentive structures can be developed, laying the most fertile ground for an ecosystem conducive to tech progress.
• The creation of more talent-development programs, similar to the program implemented by Arizona State University, whose online programs on the “Fundamentals of Microelectronics” has reached 21,000 students and is hoping to reach 50,000 by the end of 2025.
• Engage with existing automation technologies and infrastructures, such as the Japan-based Semiconductor Assembly Test Automation and Standardization Research Association (SATAS), which is developing proprietary technologies and processes for full automation of the assembly, testing, and packaging steps.³⁴
• Organizing new “Peorias”—that is, deliberately facilitating partnership agreements that involve knowledge-sharing and coordinated inter-firm procurement. These agreements not only create incentives for firms to invest in the U.S., but also enhance the technical and market competitiveness of firms that enter the U.S. ATP ecosystem.

Movement is Life: Integrating Adaptation into Policy

As long as the U.S. government continues to compete with China for technological leadership, it must strengthen both research and development, as well as its broader techno-industrial strategy. DeepSeek’s latest success suggests that the gap in innovation may not be as vast as previously assumed, reinforcing the need for continuous advancement.³⁵

The outlined strategies represent just one component of a much larger, coordinated effort required across emerging industries. Similar to Amkor’s investment in Peoria, achieving technological dominance—and, by extension, national security—depends on collaboration between industry and government to foster an environment conducive to innovation.

While a single, unified strategy may be unrealistic given the complexity of the industry, one certainty remains: bottlenecks will continue to emerge as computing technologies evolve. Addressing these challenges proactively must remain a priority.

Acknowledgments

This paper was researched and written as part of the Next Frontier Seminar, a collaborative forum for technology and policy innovation. Special thanks to the seminar’s leadership and participants for their invaluable insights. The authors extend their gratitude to Daniel Kroth, Graeme Clements, Jake Birdwell, and Joyce Guo for their leadership and guidance throughout the seminar discussions.

For more information about the Next Frontier Seminar, visit: nextfrontierseminar.org.

1. “Readout of White House CEO Summit on Semiconductor and Supply Chain Resilience,” Statements and Releases, The White House, April 12, 2021. ↩︎
2. Council of Economic Advisers, “Issue Brief: Supply Chain Resilience,” The White House, November 30, 2023. ↩︎
3. “The Global Semiconductor Supply Chain: Four Phases,” Business Executives for National Security. ↩︎
4. Semiconductor Industry Association, Beyond Borders: The Global Semiconductor Value Chain, September, 2016. ↩︎
5. National Institute of Standards and Technology, “National Advanced Packaging Manufacturing Program,” NIST, U.S. Department of Commerce. ↩︎
6. The White House, “Another Historic Investment Secured Under President Trump.” The White House, March 3, 2025. ↩︎
7. John VerWey, “Betting the House: Leveraging the CHIPS and Science Act to Increase U.S. Microelectronics Supply Chain Resilience,” Center for Security and Emerging Technology, January 2023. ↩︎
8. Semiconductor Industry Association, “Turning the Tide for Semiconductor Manufacturing in the U.S.” SIA, September, 2020. ↩︎
9. Semiconductor Industry Association, “CHIPS Incentives Awards,” NIST, January 22, 2025. ↩︎
10. Akhil Thadani and Gregory C. Allen, “Mapping the Semiconductor Supply Chain: The Critical Role of the Indo-Pacific Region,” CSIS, May 30, 2023. ↩︎
11. “Soft Launch of Penang Silicon Design 5km: Formation of Governance Structure & Incentive Packages,” Digital Malaysia, February 22, 2025; “Intel to Invest $7 Billion in New Plant in Malaysia, Creating 9,000 Jobs,” Reuters, December 16, 2021. ↩︎
12. “Vietnam Expands Chip Packaging Footprint as Investors Reduce China Links,” Reuters, November 12, 2024; Raj Varadarajan, et al., “Emerging Resilience in the Semiconductor Supply Chain,” Semiconductor Industry Association, May, 2024. ↩︎
13. “Amkor Technology Strengthens Investment in Vietnam,” Amkor Technology, October 2, 2023. ↩︎
14. “Micron Announces New Semiconductor Assembly and Test Facility,” Micron Technology, September 20, 2023. ↩︎
15. “IDC Forecasts Strong Growth in Asia/Pacific Semiconductor Market Through 2027,” International Data Corporation (IDC), December 18, 2023. ↩︎
16. U.S. Department of Commerce, “Biden-Harris Administration Announces Preliminary Terms with Amkor,” July 2024. ↩︎
17. Adam Clark, “TSMC Stock Surges After CEO Says AI Chip Demand Is Booming. It’s Good News for Nvidia,” Barron’s, October 19, 2024. ↩︎
18. “Amkor and TSMC Expand Partnership and Collaborate on Advanced Packaging,” Amkor Technology, October 3, 2024. ↩︎
19. Ibid. ↩︎
20. “High-Tech Industry in Latin America: Consumer Electronics & Semiconductors,” Americas Market Intelligence, 2024. ↩︎
21. “Anuncios de Inversión: 30 Noviembre 2023,” Gobierno de México, December 6, 2023. ↩︎
22. “Mexico’s Microchip Advantage,” Foreign Affairs, 2024. ↩︎
23. Ibid. ↩︎
24. “Foxconn Says It Is Building the World’s Largest Manufacturing Facility for Nvidia’s AI Chips,” Reuters, October 8, 2024. ↩︎
25. “NVIDIA Blackwell Platform Arrives to Power a New Era of Computing,” NVIDIA Newsroom, March 18, 2024. ↩︎
26. “ISE Announces New Site in Mexico,” ASE Global, 2024; “Micron to Build Engineering and Operations Center in Mexico,” Ventas de Seguridad, 2024. ↩︎
27. “U.S. Department of State and Inter-American Development Bank Collaborate on Semiconductor Initiative in the Western Hemisphere,” U.S. Embassy in Panama, 2024. ↩︎
28. “Roadmap for Nearshoring Investment in Mexico’s Chip Sector,” Mexico Business News, 2024. ↩︎
29. “Launch of Americas Partnership for Economic Prosperity’s Semiconductor Workforce Symposium: Expanding the Supplier Ecosystem,” U.S. Embassy in Panama, 2024. ↩︎
30. Aida Pelaez-Fernandez, “Mexico Will Not Go Tit-for-Tat on Tariffs with US, Sheinbaum Says,” Reuters, April 2, 2025. ↩︎
31. Aida Pelaez-Fernandez and Raul Cortes, “Mexico Celebrates Preferential Treatment under USMCA after US Tariffs Announcement,” Reuters, April 3, 2025. ↩︎
32. Semiconductor Industry Association, “Emerging Resilience in the Semiconductor Supply Chain,” May 2024. ↩︎
33. Less difficult than processes needed for packaging chips in AI applications. ↩︎
34. “Yamaha Motor Develops SATAS—High-Speed Semiconductor Inspection System,” Yamaha Motor Co., Ltd., May 7, 2024. ↩︎
35. Yasir Atalan, “DeepSeek’s Latest Breakthrough Is Redefining AI Race,” Center for Strategic and International Studies, February 3, 2025. ↩︎