美国应如何投资半导体的未来（英）-外交政策研究所-2021.3-26页_1mb

Summary of "LABS OVER FABS: HOW THE U.S. SHOULD INVEST IN THE FUTURE OF SEMICONDUCTORS"

Core Content

This report discusses the U.S. semiconductor industry's challenges in the face of growing global competition, particularly from China, and outlines a strategic approach to maintaining U.S. leadership in the field. It argues that the U.S. should focus not just on manufacturing, but on research, education, and open-source technologies to secure a competitive edge in the semiconductor ecosystem.

Main Viewpoints

• The U.S. semiconductor industry is at risk of losing its leadership due to China's aggressive state subsidies and technological advancements.
• Manufacturing is a complex and costly process, and while the U.S. has a strong position in chip design and EDA software, it is losing ground in fabrication and testing.
• Protectionist policies from the past, such as those used against Japanese competitors in the 1980s, have not proven effective in the face of rapid technological change and global competition.
• Innovation and business model development are more crucial than subsidies for maintaining U.S. leadership.
• Open-source chip architectures, such as RISC-V, offer opportunities for new innovation and global competitiveness, rather than being a threat to existing intellectual property.

Key Information

U.S. Semiconductor Industry Overview

• The U.S. leads in chip design and EDA software, with companies like Qualcomm, AMD, Apple, and Google playing significant roles.
• Manufacturing is largely outsourced to South Korea and Taiwan, with Intel falling behind TSMC and Samsung in precision and efficiency.
• The U.S. only produces 12% of the world’s chips, down from over a third in 1990.
• Testing and packaging is also outsourced, though this trend is beginning to reverse.

China's Semiconductor Challenge

• China's Made in China 2025 strategy aims to reduce foreign reliance and increase self-sufficiency in semiconductors.
• The Chinese government has invested billions of dollars in state subsidies for chip manufacturing and design.
• Despite these efforts, China still cannot produce the most advanced chips with the smallest feature sizes.
• The U.S. has imposed export controls on Huawei and SMIC, but China's increasing capabilities pose a long-term threat.

Historical Lessons

• In the 1980s, Japanese firms challenged U.S. dominance in DRAM production by offering higher quality and lower prices.
• U.S. protectionist measures, such as tariffs and import restrictions, failed to restore competitiveness and led to market shifts.
• Intel and EDA software firms succeeded through innovation and ecosystem building, not through government intervention.

Policy Recommendations

1. Pushing the Boundaries of Chips

   • Establish 25 university-industry centers of excellence with a focus on cutting-edge R&D.
   • Support next-generation semiconductor researchers through a national talent program.

2. Making Chip Careers Competitive

   • Create high school centers of hardware excellence to inspire and educate young students.
   • Fund a national talent program for 5,000 undergraduates and 1,000 graduate students annually.
   • Expand funding for scholars in semiconductor-related fields to promote cutting-edge innovation.

3. Opening American Semiconductors

   • Encourage open standards in defense supply chains.
   • Lead global efforts to standardize open-source models through NIST and other agencies.

Conclusion

The U.S. must shift its focus from manufacturing to research and innovation to maintain its leadership in the semiconductor industry. Subsidizing manufacturing is not the most effective way to secure long-term advantages. Instead, the U.S. should invest in education, support open-source technologies, and foster a robust and diverse semiconductor ecosystem. By doing so, it can ensure continued global leadership in the field and drive future innovation.