The US State Department says it has "serious concerns" about American companies using AI systems made in China, arguing they are designed to promote China's views and limit free speech. The warning reflects growing alarm among US lawmakers who want to slow this trend.

The shift is happening at major companies, not just small startups. Coinbase, the cryptocurrency exchange, disclosed it uses two Chinese AI systems. Other large firms like Airbnb and Uber are doing the same, according to recent reporting. One startup called Lindy switched to a Chinese AI model called DeepSeek in June because it cost much less to run.

Two main reasons explain why this is happening so fast. First, money: Chinese AI systems cost significantly less to use than American ones like OpenAI and Anthropic. As these American companies raise their prices, US firms are looking for cheaper alternatives. Second, access: many Chinese AI models are published as open-source software, meaning any company can download the code and run it on its own computers instead of paying to use it through a website. This makes Chinese AI easy to get.

Here's where policy gets tricky. The US government has already restricted some American AI models at home, asking companies to stop using certain frontier systems. But instead of stopping AI use altogether, this seems to have pushed some companies to switch to Chinese alternatives. US officials worry this gives Chinese AI companies time to catch up and improve their technology.

Over the past year, Washington has been building rules around AI. The White House released plans in July 2025, updated policies in December 2025 and March 2026, and issued executive orders in April and June 2026. The goal is to keep American AI advanced and safe. The US has also asked allied countries to prevent overseas companies from accessing the most powerful American AI systems.

At the same time, the government is blocking China from getting advanced computer chips used to build AI. In May, the US stopped allowing Nvidia to ship its best chips to Chinese companies. The US also restricted investment in Chinese AI companies, and China responded by planning similar limits on US investment in Chinese firms.

China has changed its approach too. For years, Chinese AI companies offered their systems free or cheap worldwide to gain users. Now Beijing is considering blocking its best AI systems from being used overseas. Chinese regulators have also accused an American AI company, Anthropic, of hiding security flaws in its code.

Both countries are now trying to keep each other's technology out. But here's the problem: companies still want to use cheaper or better tools, no matter where they come from. Cost savings and easy access to open-source code are powerful forces. Right now, what companies choose to buy is moving faster than what the government can regulate.