Information source:
https://www.scmp.com/news/china/science/article/3325106/white-house-adviser-calls-us-defend-its-ai-lead-against-chinese-advances?module=top_story&pgtype=section

Michael Cracios, director of the White House Office of Science and Technology Policy, warned at a Senate hearing that the United States' absolute advantage in the field of artificial intelligence is "significantly narrowing", calling for the maintenance of the United States' technological hegemony by exporting a complete "AI technology stack". This strategic shift marks the United States' shift from traditional chip export control to full-stack technology output, trying to curb China's technological rise by building a US-centric AI ecosystem. Michael Krasiostells the U.S. Senate subcommittee that exporting a "cohesive AI technology stack" is key

The narrowing of the technology gap has caused US strategic anxiety

The data revealed by Clasios at a Senate subcommittee hearing deeply concerns U.S. policy makers. In 2020, the United States had remained far ahead of its closest competitors, including China, in the field of AI, but by 2024, this technology gap has begun to "significantly narrow." This change prompted the U.S. government to re-examine its competitive strategy in the AI ​​field.

According to the analysis reports of several US think tanks, China has accounted for 70% of the world's share of AI patent applications, and has even surpassed the United States in certain specific technical fields. Especially in key technical directions such as large language models, computer vision and natural language processing, Chinese companies have shown strong innovation capabilities. The technological breakthroughs of Chinese AI companies such as DeepSeek have made the United States feel unprecedented competitive pressure.

The report "Key and Emerging Technologies Index" released by the Center for Strategic and International Studies shows that although the United States generally maintains a leading position in the field of AI, China has made significant progress and has unique advantages, and is expected to pose a substantial challenge to the US AI hegemony in the next decade. This technical catch-up situation forced the US government to shift from passive defense to proactive attacks.

The Trump administration clearly listed the confrontation with China as one of its core goals in the 28-page "Win the AI ​​Competition: The US AI Action Plan" released in July this year. The plan covers more than 90 policy recommendations, from relaxing supervision, accelerating infrastructure construction to promoting technology exports, forming a comprehensive competitive strategy. Unlike the "high wall" strategy that the Biden administration mainly relies on chip bans, the Trump administration is more inclined to build a global AI ecosystem dominated by the United States through technology output.

Deep logic of technology stack output strategy

The concept of "AI Technology Stack" represents an important transformation in the US AI competitive strategy. Traditional technology stacks usually include advanced chips at the hardware level, operating systems and middleware at the system level, and algorithms and software at the application level. Cracios stressed that what the United States needs to output is a "cohesive AI technology stack", that is, a complete technical system covering the bottom-level hardware to the upper-level applications.

The core logic of this strategy is to maintain the competitive advantage of the United States through technological dependence. If the development of AI in other countries is based on the technology stack provided by the United States, then the United States can influence the global AI industry structure through technological control at critical moments. As Cracios said, "To have a cohesive, successful AI ecosystem, you must have physical computing power to run large language models and applications built on top of them."

At the same time, this strategy also reflects the US's doubts about the effect of pure export controls. Although the chip ban in the past few years has restricted China from obtaining advanced chips in the short term, it has also prompted China to increase its investment in independent research and development and achieve technological breakthroughs in some areas. The technology stack output strategy attempts to restrict opponents from obtaining key technologies through the "carrot and stick" method, and maintain the US's technological influence through technical cooperation.

The U.S. Department of Commerce will work with the White House Office of Science and Technology Policy to develop specific details of the export of technology stacks, which Krasios calls "one of the most important actions" in the U.S. AI action plan. It is expected that the relevant policies will involve export licensing, technology transfer restrictions, and coordination mechanisms with allied countries.

Challenges and opportunities for reshaping the global AI industry structure

The implementation of the US technology stack output strategy will have a profound impact on the global AI industry structure. On the one hand, countries that rely on American technology may acquire more advanced AI capabilities and accelerate their digital transformation process; on the other hand, this may also intensify global technological differentiation and form two parallel technology ecology with the United States and China as the core.

For China, the new strategy of the United States is both a challenge and an opportunity. The challenge is that if the United States successfully builds a global AI ecosystem with its technology as its core, China may face greater pressure on technology isolation. But the opportunity lies in that this pressure will further promote China to accelerate its independent innovation and achieve "overtaking on the curve" in some technical fields.

China has shown strong independent innovation capabilities in the field of AI. From Baidu's Wen Xin Yiyan to Alibaba's Tongyi Qianwen, from Huawei's Ascend chips to Cambrian's AI processors, Chinese companies are building a relatively complete AI technology ecosystem. Especially in application innovation, Chinese companies are often able to transform AI technology into actual business value faster.

Other major economies such as the EU are facing pressure to take sides in the technological competition between China and the United States. Although the EU is at the forefront in formulating AI regulatory frameworks, it is relatively weak in core technology capabilities and relies more on the US technology supply. How to safeguard technological sovereignty while avoiding marginalization will be an important challenge facing the EU.

The U.S. government also instructed the Department of Defense and the Department of Commerce to coordinate with their allies to adopt export control measures and exclude "adversaries" such as China from the U.S. defense supply chain. This "circle of friends" strategy aims to amplify the US technological influence through cooperation with allies, but may also trigger countermeasures from other countries.

In the long run, the global AI industry may form a multi-polar competitive landscape. The United States still maintains its advantages in certain core technology fields, China has unique advantages in applied innovation and manufacturing capabilities, the EU plays an important role in setting regulatory standards, and other emerging economies may find their own technological paths in specific segments. Although this multipolar competitive landscape will increase the complexity of technological coordination, it may also inject more innovative vitality into global AI development.