The United States hasn't announced a blanket ban on Chinese AI models. But officials have considered sanctions and other restrictions against Chinese developers over concerns about intellectual property, access to advanced chips, and the security of powerful open-weight models. 

The immediate dispute has centred on Moonshot AI, the company behind Kimi K3. Broader action could also affect how American cloud providers, AI-routing platforms and businesses handle models from DeepSeek, Alibaba’s Qwen and other Chinese developers. 

Chinese models already account for about 63% of token usage by US companies on OpenRouter, a platform that lets developers access multiple AI models through one API, according to Yahoo Finance. That does not mean Chinese models represent a significant portion of the entire American AI market. But it shows how important they have become to developers using the platform. 

Techloy therefore ran a switching-cost stress test: if a business lost access to its current model and moved the same monthly token workload to another provider, how would its API bill change? 

Is it safe to use AI chatbots? I tried jailbreaking
Over two weeks, I tested ChatGPT, Claude, Gemini, Grok and DeepSeek using the same jailbreak techniques to see which models held up and which didn’t.

Subscribe for free to continue reading this article

Subscribe Subscribe

Already Have an Account? Log In