Recent Ban on Manus Acquisition

Recently, the Office of Foreign Investment Security Review in China (National Development and Reform Commission) legally prohibited a foreign investment in the Manus project, requiring the parties involved to withdraw from the acquisition. Manus is an intelligent product released by Butterfly Effect Company in March 2025. In December 2025, US company Meta announced plans to acquire Manus for approximately $2 billion. The core business of this AI company involves a complex transfer of personnel, technology, and data from China to overseas. According to the Foreign Investment Security Review Measures, such investment activities in technology must undergo a security review.

This incident serves as a significant warning: technology knows no borders, and the political nature of the competition surrounding technological research and related capital relationships is particularly strong, especially as AI becomes a core area of competition among major powers.

The US Government’s All-in Strategy on AI

The Trump administration viewed technological competition as a core aspect of US-China rivalry, emphasizing advanced semiconductors, artificial intelligence, autonomous systems, biomedicine innovation, and cybersecurity. The US is concerned about China’s growing innovation capabilities. The Information Technology and Innovation Foundation released a report stating that China is at or near a global leading level in nuclear power, robotics, AI, and quantum computing. Robert Atkinson, the foundation’s president, noted that China’s science and technology system has a “real advantage,” asserting that China is not a small dragon but a fire-breathing giant, potentially catching up to or surpassing the US and other advanced Western countries within a decade.

After Trump began his second term, he quickly adjusted US AI policy. On January 23, 2025, he signed an executive order to eliminate barriers to US AI leadership, rolling back many regulatory measures from the Biden era. The Trump administration exhibited a “technological accelerationism” approach, aligning with the interests of Silicon Valley’s tech and capital groups.

During Biden’s term, the US government focused on managing risks associated with AI technology development, while the Trump administration prioritized “accelerating innovation.” Traditionally, Trump communicated with Michael Kratsios, the director of the White House Office of Science and Technology Policy, in March 2025, stating that the US must prioritize the development of AI, quantum information science, and other key emerging technologies to maintain global leadership. In this letter, Trump specifically mentioned China’s rapid development in AI technology.

In early December 2025, the White House released the National Security Strategy report for Trump’s second term, making significant adjustments to the definition of US national security interests and foreign strategic layout, reflecting the administration’s policy focus on “strength and profit maximization.” The report stated, “The US must maintain the strongest economy, develop the most advanced technology, enhance our cultural health, and deploy the world’s most combat-capable military.” In the technology policy section, the new national security strategy emphasized the need to maintain and expand US advantages in cutting-edge military and dual-use technologies, particularly in AI, quantum computing, and autonomous systems. Additionally, the Trump administration proposed strengthening the resilience of the US tech sector and encouraging high-tech cooperation to align the policies of related countries in favor of the US.

From the US perspective, AI technology is fundamentally changing the underlying logic of great power competition. As a foundational and general-purpose technology, AI is a field the US must dominate. In July 2025, the White House released the “Winning the Race: America’s AI Action Plan,” outlining arrangements to enhance collaboration in computing power, data, electricity, and talent, while emphasizing increasing the full-stack export of US AI technology. This strategic report was personally released by Trump, who clearly viewed AI leadership as a competition based on “global technological dominance, market share, and rule-making power,” with “winning against China” as a core policy objective. On the same day this document was released, Trump signed three presidential executive orders to comprehensively deploy around the three pillars of “accelerating innovation, building infrastructure, and leading international diplomacy and security,” aiming to construct a framework for advancing AI technology and industry in the US.

The Trump administration also focused on leveraging AI to promote innovation in energy, critical materials, and advanced manufacturing. In November 2025, Trump signed an executive order to launch the “Genesis Mission,” which aims to fully integrate the US computing infrastructure, promote the fusion of critical datasets, and establish mechanisms like the “American Science and Security Platform” (ASSP) based on AI technology, facilitating breakthroughs in controllable nuclear fusion, quantum information, and biomanufacturing.

This initiative is highly task-oriented and can be seen as the US’s “national system” in the tech sector. The project is led by Dario Gil, the Deputy Secretary of Energy, who previously headed research at IBM. Gil described the “Genesis Mission” as a “decisive moment for the next era of American science.” The Department of Energy plays a crucial role in the US tech innovation ecosystem, managing over ten of the country’s top national laboratories and providing substantial research funding to US universities and research institutions.

From the perspective of competing for the global AI commercial ecosystem, the Trump administration has also introduced numerous initiatives. In July 2025, the US released the “Winning the Race: AI Action Plan,” proposing to export full-stack US AI technology to allies and partners. The term “full-stack” reflects that AI is essentially a technology stack, comprising various core elements such as infrastructure components, models, and software. In other words, the US aims to export an “AI technology package” to relevant countries and regions, including Nvidia’s GPUs, Microsoft’s Azure, Amazon’s AWS cloud services, as well as OpenAI and Anthropic’s large models and API interfaces.

To promote the global adoption of US AI technology stacks, the US Department of Commerce encourages related companies to form coalitions and provides financial support for this “packaged export” of AI technology. Michael Kratsios, the director of the White House Office of Science and Technology Policy, stated in a congressional hearing that full-stack technology export is crucial for the US to seize the AI technology lead, noting, “When China has the capability to export chips on a large scale, the US will already be widespread globally.”

The Changing Logic of US-China AI Competition

Although there is still a gap compared to the US, China has become a leader in the AI field. On April 13, Stanford University’s Human-Centered AI Institute released the “2026 AI Index Report.” According to this report, the performance gap between AI models from the US and China is narrowing. From early 2025 to March 2026, the top models from both countries alternated leading positions on the Arena leaderboard. In February 2025, DeepSeek-R1 nearly matched the best US model; by March 2026, the top US model Claude Opus 4.6 (1503 points) only led China’s top model Dola-Seed-2.0 Preview (1464 points) by 2.7 percentage points.

The US still holds advantages in private sector investment, the number of top models, high-impact patents, and chip and computing power infrastructure. In 2025, the US released 50 advanced AI models of interest, while China had 30. The US has 5,427 data centers, and its advantage in computing power infrastructure largely stems from significant capital expenditures in the private sector.

China, on the other hand, leads in the number of published papers, citations, and patent outputs, especially in the deployment of industrial robots (physical AI), where it ranks first in the world. In September 2025, the International Federation of Robotics released the “2025 World Robotics Report,” showing that in 2024, China’s stock of operational industrial robots surpassed 2 million units, ranking first globally. Since 2017, Chinese factories have installed more than 150,000 robots annually. China’s share of the global robotics manufacturing industry has risen to 33%, up from 25% in 2023. AI-driven humanoid robots have become a new highlight of China’s economic development, aiding the transformation and upgrading of Chinese manufacturing while also promoting the iteration of AI technology and the practical application of specific scenarios. The deep integration of technology and manufacturing is a prominent feature of China’s AI industry development.

Interestingly, just one month after the Trump administration released the “Winning the Race: AI Action Plan,” the State Council of China issued the “Opinions on Deepening the Implementation of the ‘AI+’ Action Plan” in August 2025, proposing goals such as achieving over 90% penetration rates for new-generation intelligent terminals and intelligent agents by 2030.

It is evident that the focus of the US-China tech competition is shifting from accelerating technological breakthroughs to building innovation ecosystems and industrial systems. The Trump administration’s approach of directly investing in companies like Intel reflects an attempt to strengthen US technological superiority through “state capitalism,” addressing shortcomings in advanced manufacturing. The Trump administration also aims to use US financial tools to restrict Chinese tech companies globally.

Recently, the US Congress has proposed a series of bills to strengthen export controls on China in areas such as cloud computing power and lithography machines, advocating that allies like the Netherlands and Japan should adopt similarly strict measures against China to fill the so-called loopholes in the US’s tech war against China.

In this context, China is taking multiple measures to accelerate high-level technological self-reliance, deepen the integration of production and innovation, and strive to maintain an international tech cooperation network, enhancing the overall effectiveness of its national innovation system. Meanwhile, China is increasingly establishing a more detailed governance system for economic and technological security. The recently announced national “14th Five-Year Plan” emphasizes that coordinating development and security is a major principle that must be followed in promoting economic and social development.

On April 20, Zheng Shanjie, the director of the National Development and Reform Commission, stated in an article in the People’s Daily that continuously enhancing national economic security capabilities is a key task during the “14th Five-Year Plan” period. In analyzing the challenges facing China, he emphasized that the acquisition of technology in certain key areas is restricted, and there are still “bottleneck” risks in critical components, key basic materials, and high-end equipment. The safety regulation of new technologies and new business formats, such as AI, still needs to be improved, and risks such as data leakage, deep forgery, and cyberattacks cannot be ignored. Zheng stated that efforts will be made to promote the construction of security capabilities in model algorithms, data resources, infrastructure, and application systems, deepen international cooperation in AI, and effectively safeguard national security interests in the context of opening up.

Clearly, the Manus incident serves as a warning for companies in related fields. In the context of the US’s ongoing “tech war” against China, particularly the intensifying competition in AI, technology companies must deeply understand the complexities of geopolitical competition and remain vigilant about the security risks inherent in related technology development activities and capital interactions. It is crucial to recognize that during Trump’s second term, the US is likely to intensify its tech competition with China, imposing stricter external barriers on China’s access to key technologies. The US’s ongoing intention to build a “tech alliance” against China is also evident. The research and application of AI technology are rapidly changing the underlying logic of national security and great power competition, and relevant companies and investment institutions need to learn lessons from the Manus incident. Balancing security and development is not only a major concern for the government but also a critical question for market participants.