The Trump administration has issued a de facto ultimatum to more than 30 countries participating in U.S.-led AI cooperation frameworks, demanding they cease parallel participation in Chinese-led initiatives. This strategic move is intended to thoroughly protect intellectual property and eliminate China from advanced technology supply chains to secure absolute American dominance in next-generation AI.

Global Demand for Allegiance in the AI Era

Loyalty to the U.S.-Led “Pax Silica”

In August 2026, the Trump administration began preparing formal letters to notify over 30 countries involved in the U.S.-led “Pax Silica” framework that parallel participation in Chinese-led AI organizations will no longer be tolerated. A draft of the letter from the U.S. State Department takes an uncompromising tone, stating that “participating in everything is the same as participating in nothing” and demanding that nations make a “deliberate choice” regarding their AI alignment. This policy directly impacts the 35 nations and regions—including Japan, South Korea, and the European Union—that signed the “AI Opportunity Statement” in June 2026. The U.S. government defines Pax Silica not merely as a membership group, but as a binding commitment to reduce reliance on adversarial nations for semiconductors and critical minerals. Currently, Kazakhstan is the only confirmed mineral-supplying nation participating in both U.S. and Chinese frameworks, a “double-dipping” status that has raised serious alarms in Washington.

Conflict with China’s “World AI Cooperation Organization”

In direct opposition to the American initiative, Chinese President Xi Jinping announced the establishment of the “World AI Cooperation Organization” in Shanghai in July 2025. To date, 29 countries, including Russia, Brazil, South Africa, and Pakistan, have signed the founding agreement. China aims to expand its global influence by promoting “open-weight” AI technology, which allows for broader modification and distribution compared to proprietary U.S. systems. U.S. officials have expressed that it is increasingly difficult for any nation to be viewed as a “trusted partner” within the Western tech ecosystem while simultaneously signing onto initiatives designed to advance China’s competing vision of AI. This movement toward alignment is not just a political gesture but a physical partitioning of the technology sphere, extending from advanced AI chips and models down to the raw materials used in their production. The following chart illustrates the current global alignment between these two competing AI camps.

Protecting AI Technology and Countering “Model Distillation”

An Unprecedented Alliance of U.S. AI Giants

To prevent the leak of critical AI technology, major U.S. companies including OpenAI, Anthropic, and Google have formed an unprecedented alliance to counter intellectual property theft by Chinese startups. At the center of this conflict is a technique known as “model distillation,” where Chinese companies allegedly use the outputs of advanced U.S. AI models to train their own smaller, cheaper models without authorization. This allows competitors to replicate high-end capabilities at a fraction of the original development cost—a practice U.S. tech leaders label as “free-riding” on American R&D. The “DeepSeek Shock” in early 2025, sparked by a Chinese startup releasing the R1 model with performance rivaling U.S. counterparts at significantly lower costs, served as a massive wake-up call for the industry. In response, these American companies are now sharing defensive strategies through the Frontier Model Forum to detect and block these covert “distillation attacks”.

Software Distribution Bans and Security Risks

The scope of U.S. regulations is shifting from hardware (such as chips) to the very distribution of AI software and models. A primary concern for the Trump administration is the proliferation of “open-weight models,” such as Meta’s Llama series or Alibaba’s Qwen, which serve as foundational building blocks for developers worldwide but are nearly impossible to control once released. The U.S. government now views AI models with advanced hacking or weapon-development capabilities as sovereign security threats and is considering strict access controls for foreign nationals. A notable example occurred in June 2026 when Anthropic was ordered to halt the provision of its latest “Claude Mythos 5” model to non-U.S. citizens due to its “uncanny knack” for finding software vulnerabilities. This incident has exposed the inherent vulnerability of a global society increasingly dependent on centralized AI that can be revoked by a single government’s order. The diagram below outlines the mechanism of distillation attacks and how they pose a risk to U.S. national security.

Purging Supply Chains and the Resilience of Chinese Enterprises

The “Purge” from Advanced Tech Supply Chains

The U.S. government is accelerating measures to completely exclude China from advanced industrial supply chains, including components for AI data centers, robotics, and semiconductor materials. The Federal Communications Commission (FCC) in the United States is drafting a bill to ban the import of Chinese-made “optical transceivers,” which are essential for transmitting information between data centers, citing risks of data theft or service disruption. This measure is expected to heavily impact companies like Zhongji Innolight, which holds a 27% share of the global market. Furthermore, the administration is preparing a “price floor” and high tariffs on polysilicon, a key material for both solar panels and semiconductors, to protect the domestic U.S. industry and check Chinese dominance. Import bans are also being discussed for humanoid and quadrupedal robots, signaling that the U.S. intends to sever ties in the realm of “Physical AI” where intelligence meets hardware.

Adaptation of Chinese Firms in the “Chip War”

While U.S. sanctions have dealt significant blows, they have also inadvertently accelerated technological self-reliance within China. For instance, Huawei, despite being cut off from cutting-edge global chip supplies, successfully developed its own “Kirin” processors and “HarmonyOS” operating system with massive government backing. As of 2024, HarmonyOS has surpassed Apple’s iOS in market share within China. Similarly, SMIC (Semiconductor Manufacturing International Corporation), China’s largest foundry, has pushed forward with the mass production of 14nm and 7nm process nodes despite restrictions on advanced lithography equipment from the West. The “Chip War” has thus evolved from a series of one-sided sanctions into a long-term endurance race where both the U.S. and China are building entirely independent and complete technology ecosystems.

Future Outlook

The primary focus moving forward is whether the U.S. demand for side-selection will lead to a “Technical Iron Curtain” that permanently divides the global AI landscape. If restrictions on open-weight models and software collaboration become a permanent reality, it could fragment the global research community and lead to a stagnation of innovation due to fractured standards. Policymakers worldwide, including those in Japan, now face the difficult task of weighing the benefits of access to the secure, U.S.-led supply chain against the cost of losing access to the vast Chinese market and its low-cost AI advancements.

Reference Pages

• 【International Journal of Social Sciences and Public Administration】https://ijsspa.org

• 【Built In】https://builtin.com/articles/openai-google-anthropic-ai-model-theft-china

【#AIRace #TrumpAdministration #PaxSilica #TechHegemony #Decoupling #IntellectualProperty #SupplyChain #DeepSeek】