Trump must stop China’s AI intellectual property theft

The race between the United States and China for superiority in artificial intelligence has suddenly become far more competitive. The challenge? China’s use of “distillation” to introduce new, lower-cost, open-weight AI models that are highly competitive with the most advanced U.S. AI models.

Distillation refers to using the output of a frontier AI model to train another model. It is like a student who never attends class asking, “Can I copy your homework?”

On July 17, the Chinese AI company Moonshot introduced Kimi K3, arguably the world’s largest open-weight AI model. Analysts say Kimi K3 delivers performance approaching that of the most advanced models produced by Anthropic and OpenAI, two of America’s leading AI companies. Anthropic, OpenAI, Alphabet’s Google, and other American firms spend hundreds of millions of dollars developing so-called frontier models that push the boundaries of AI. Chinese AI companies such as Moonshot then spend only a fraction of that amount by distilling these American innovations and copying, adapting, and refining before releasing lower-cost models to the global market.

President Donald Trump is now debating how to respond to this competitive challenge. Jensen Huang, the chief executive officer of Nvidia, along with several other leading AI executives, recently argued in a widely publicized post on X that the U.S. should welcome Chinese AI models because they expand knowledge and competition, both of which ultimately benefit AI development globally.

Huang and the other American AI leaders are mistaken. The U.S. should insist that China and its AI companies conform to accepted principles of economic competition. Intellectual property theft cannot be tolerated. Nor should China be allowed to use distilled AI models to advance its broader mercantilist trade agenda. China heavily subsidizes AI development, enabling its companies to price models below cost in an effort to drive American competitors out of global markets. AI would become yet another instrument for expanding Chinese economic and geopolitical influence.

On a level playing field, the U.S. can and will remain ahead of China in developing frontier AI models. The U.S. enjoys commanding advantages in the design of accelerated computing platforms and application-specific semiconductors. Through targeted export controls, the U.S. denies China access to the most advanced electronic design automation software and critical chip manufacturing equipment, including ASML’s extreme ultraviolet lithography systems. In addition, the world’s most advanced American-designed chips are fabricated by Taiwan Semiconductor Manufacturing Company, whose manufacturing technology remains years ahead of China’s most sophisticated semiconductor facilities.

America’s cloud providers also possess a 5-1 advantage over China in total computing capacity. That allows U.S. companies to conduct larger mathematical experiments, generate far more training data, and deploy massive inference models. Finally, the U.S. enjoys a decisive advantage in financing AI development. Between 2026 and 2028, American companies are expected to invest more than $3 trillion in AI models and the infrastructure needed to deploy them worldwide. China cannot match that level of capital investment. America’s deep and liquid capital markets provide an enduring structural advantage, allowing U.S. companies to continue investing for as long as necessary to maintain leadership in AI.

The U.S. will remain the global leader in AI because American companies compete relentlessly against one another. Competition, not government direction, determines which products succeed. American AI companies invest heavily in research, constantly innovate, and strive to build the world’s most advanced models. They understand that sustained leadership comes from producing superior AI systems at market-driven prices, not from copying the work of others.

HOW THE BIG BEAUTIFUL BILL BECAME LAW: READ MORE FROM THE WASHINGTON EXAMINER’S 10-PART SERIES

That competitive culture remains America’s greatest advantage in the race with China for AI superiority.

The writer owns shares in Nvidia and Alphabet.

James Rogan is a former diplomat who later worked in law and finance for over 30 years. He writes a daily note on markets, economics, politics, and social issues.