US Ban on Chinese AI Models Could Cost American Businesses $12B Annually

According to Daniel Yue, an assistant professor at the Scheller College of Business at Georgia Institute of Technology, a possible ban by the US on Chinese open-weight artificial intelligence models could lead to an annual financial burden of up to $12 billion for American companies. This estimation stems from usage statistics provided by OpenRouter, a New York-based aggregator of large language models (LLMs), which allows developers to utilize various AI models via a single API. Yue determined that if users of OpenRouter had to switch from Chinese models to leading proprietary options, they might incur an extra $2 billion annually, based on token usage and pricing discrepancies noted from July 21 to 27. When applied to the wider US economy, the increased costs could vary from $3 billion to $12 billion, depending on the nation's dependence on these Chinese models. Yue emphasized that these figures serve as a rough estimate rather than an exact forecast due to the challenges in monitoring usage beyond centralized platforms. Reports indicate that Washington is contemplating a ban on foreign open-source AI models, which would likely elevate expenses for numerous US businesses that are increasingly dependent on affordable Chinese alternatives. While the precise economic impact is hard to gauge, the findings underscore the potential financial consequences of such a policy.