Following the launch of the Kimi K3 model developed by Moonshot AI, the US administration is again accelerating its efforts to ban China-based artificial intelligence models. According to the report published by Axios, cyber security concerns are at the heart of this attack.

The open nature of models such as DeepSeek and Kimi K3 allows users to host these models in their own infrastructures. This is widely used among US-based businesses as it helps companies maintain data privacy and reduce costs.

Details of the ban attempt and sectoral impacts

Last year, the US Department of Commerce considered adding many Chinese artificial intelligence laboratories, including DeepSeek, to the “Entity List”, which restricts access to sensitive technologies. The National Security Agency and the Office of the National Cyber ​​​​Director also planned to conduct a joint consultation process to determine the use of Chinese models.

Critics argue that such a ban would stifle innovation and leave the market confined to a few big players such as OpenAI and Anthropic. Former White House advisor Sriram Krishnan and David Sacks state that this breakthrough aims to eliminate open source competition.

The low cost advantage offered by Chinese models directly affects the preferences of developers. For example, while the DeepSeek-V4-Pro model demands a price of $0.87 per million output tokens, this number can rise to $50 for Anthropic’s Claude Fable 5 model. Companies like Coinbase significantly reduce their artificial intelligence expenses by using these models in their production processes.

Implementation challenges and future strategies

The distribution of open-loaded models as downloadable documents on the Internet makes it technically difficult to implement a centralized ban. Since companies can download these models to their local servers and run them completely offline, it becomes almost impossible for regulatory authorities to monitor local use.

Tweaks and customizations made to models make it even more complicated to track down software originating from China. It is stated that the US government may focus on dissuading companies from these models through procurement rules and public pressure, rather than a direct ban.

Government sources aim to discourage use by pointing out potential vulnerabilities and governance problems in Chinese models.

These developments are seen as a reflection of the wide-ranging trade tensions between the USA and China in the technology sector. Washington had previously imposed restrictions on critical equipment exports, while Beijing had accelerated the development of its own local technologies.

The US administration continues its aim to maintain leadership in the global artificial intelligence race.

How do you think these possible restrictions on Chinese artificial intelligence models will change the competitive stability in the technology world?