Foreign Regulations Are Undermining Western AI Competitiveness and Benefiting China

In early 2024, engineers at one of America’s largest technology companies were working on a problem that had nothing to do with advancing technology. Meta had dedicated 590,000 engineering hours to complying with the European Union’s Digital Markets Act, the equivalent of 355 full-time engineers spending the year on compliance documentation rather than product. They were not building the next frontier models or hardening AI safety systems. They were completing paperwork for a foreign regulator.

Those lost hours are a small piece of a much larger pattern. Washington has spent several years investing seriously in American AI competitiveness, such as with the CHIPS and Science Act, which directed tens of billions toward domestic semiconductor manufacturing, and expanded federal AI research funding. It has also attempted to restrain China’s AI ambitions by deploying export controls to slow China’s access to advanced chips. That logic is sound, but it has a blind spot. While the United States is building AI capacity at home, discriminatory foreign regulations from its allies are weakening it from abroad and strengthening China’s competitive position.

The United States faces a growing competitive challenge in AI from China. Chinese firms are narrowing the performance gap with U.S. models and are no longer simply copying Western innovations. The January 2025 release of DeepSeek’s R1demonstrated their ability to produce competitive models at lower cost. A 2025 study by RAND Corporation found that Chinese LLMs’ global market share surged from 3 percent to 13 percent in just two months after R1’s release, with the largest gains in developing countries. China is not just catching up technically. It is expanding its footprint in the markets where the next generation of AI infrastructure will be built.

Against that backdrop, the regulatory environment facing the companies driving Western AI development matters enormously. Foreign digital regulations are producing a measurable drag on AI investment. The EU’s Digital Markets Act designated five U.S. companies as gatekeepers while capturing only one Chinese service. GDPR enforcement follows the same pattern: U.S. companies have paid 83 percent of all GDPR fines despite representing only 10 percent of the EU’s information and communications technology service imports. In 2024 alone, the EU extracted $6.7 billion from American technology companies, an amount equivalent to nearly 20 percent of total EU tariff revenue.

And it keeps growing. The EU’s AI Act becomes fully enforceable in August 2026, with fines of up to 7 percent of global revenue. The European Commission’s first review of the DMA, due May 2026, will assess how the law is functioning and whether its scope should be expanded, and the review has already broadened to consider whether cloud computing should fall within its scope. Microsoft delayed the launch of its AI assistant in Europe due to DMA compliance uncertainty, redirecting engineering resources from product development to regulatory alignment at a moment when speed to market in AI carries significant competitive cost.

It is not just European rules weakening American tech firms. Australia, Brazil, India, Japan, South Korea, Turkey, and the United Kingdom have all proposed or enacted similar frameworks targeting the same set of U.S. firms. The result is a regulatory environment that is fragmenting the Western AI ecosystem at precisely the moment China is working to consolidate its own.

That fragmentation flows directly to China’s benefit. While leading Western AI firms divert engineering talent and capital into compliance, China’s state-backed competitors operate under no comparable burden at home and are expanding aggressively in markets where U.S. firms face the heaviest regulatory constraints. The DMA’s asymmetric gatekeeper designations effectively give Chinese platforms a structural advantage in European and third-party markets that they did not earn through competition. Regulatory fragmentation also erodes the network effects and global scale that have allowed Western AI platforms to outpace China’s state-directed model. A fragmented Western ecosystem is a weaker one.

The implications extend beyond any single country’s commercial interests. This is not primarily a bilateral dispute between the United States and its trading partners. It is a question about what kind of AI ecosystem the democratic world wants to sustain. The companies absorbing billions in fines and compliance costs are the same companies investing in frontier AI models, building the cloud infrastructure that democratic government depends on, and setting the technical standards that will shape how AI develops globally. Weakening them does not produce a more competitive European or Brazilian or Indian AI sector. It produces a vacuum that China is well-positioned to fill.

The AI competition with China will be decided by which country can sustain investment and scale over decades. Foreign regulations that divert billions in engineering capacity from American firms through fines and compliance costs are as consequential as any model release from a Chinese lab. Right now, the U.S.-led AI ecosystem faces one of its biggest threats not from its rivals but from its allies. Governments that want to remain meaningful participants in the AI ecosystem, rather than consumers of China’s model, have a shared interest in ensuring that the Western ecosystem remains strong enough to compete.