Washington, July 24: The United States has accused Chinese artificial intelligence startup Moonshot AI of unlawfully using proprietary technology from American AI company Anthropic to develop its latest large language model, intensifying the ongoing technology dispute between Washington and Beijing.
Speaking on Wednesday, Michael Kratsios, a senior technology official in the Trump administration, alleged that Moonshot AI relied on unauthorized access to Anthropic’s most advanced AI model, known as Fable, while developing its recently launched K3 large language model.
Kratsios described the alleged activity as a serious attempt to exploit American innovation, saying that large-scale industrial distillation aimed at extracting proprietary AI capabilities threatens US technological leadership and undermines years of research and development.
The allegations come amid growing concerns in Washington over the transfer of advanced artificial intelligence technologies to Chinese firms, particularly as competition intensifies in the global race to develop next-generation AI systems.
In a separate statement posted on social media platform X, US Treasury Secretary Scott Bessent said the administration is considering placing Moonshot AI on a US trade blacklist. He also indicated that additional sanctions against the company remain under review, although no final decision has been announced.
If implemented, the proposed measures could significantly restrict Moonshot AI’s access to American technologies, software and business partnerships. Inclusion on a trade blacklist would also make it more difficult for the company to acquire products and services from US suppliers.
Kratsios further claimed that Moonshot AI has obtained advanced Nvidia artificial intelligence chips, hardware that has become central to the development and training of cutting-edge AI models. The United States has tightened export controls on high-performance AI chips in recent years to limit China’s access to sensitive computing technologies.
The latest accusations are expected to further strain relations between the world’s two largest economies, where artificial intelligence has emerged as a major area of strategic competition alongside semiconductors, quantum computing and advanced manufacturing.
China swiftly rejected the allegations. Liu Chang, spokesperson for the Chinese Embassy in Washington, dismissed the claims as “entirely unfounded” and defended China’s record on intellectual property protection.
Liu said China respects intellectual property rights and urged US officials to base their statements on facts rather than prejudice. He also called on Washington to stop what he described as attempts to discredit the achievements of China’s rapidly growing artificial intelligence industry.
The dispute reflects broader geopolitical tensions surrounding AI development, with both countries investing heavily in advanced technologies viewed as critical to future economic growth and national security.
Industry analysts believe any new sanctions against Moonshot AI could have wider implications for global AI supply chains, cross-border technology partnerships and international research collaboration.
The case also highlights increasing scrutiny over how AI models are trained, particularly concerns surrounding the use of proprietary data and the legal boundaries governing model distillation and knowledge transfer.
As the Trump administration weighs further action, technology companies and investors will be closely watching whether the dispute leads to additional export restrictions or broader measures affecting the rapidly expanding global artificial intelligence sector.