Despite Trump crackdown, Americans turn to Chinese AI as cheaper models challenge ChatGPT and Claude
China’s artificial intelligence (AI) models are growing in popularity in the United States, despite the Trump administration’s tougher stance on Chinese technology. Many Americans are abandoning popular American AI tools, such as Chat GPT Claude chose Chinese AI models because they were cheaper and could do many daily tasks well.
Raffi Krikorian, chief technology officer of San Francisco-based Mozilla, said he started using Chinese artificial intelligence startup Moonshot’s Kimi K3 just days after it launched. Krikorian said he uses Kimi K3 for many daily tasks instead of Anthropic’s Claude chatbot. “It seems faster,” he said when comparing K3 to Anthropic’s Claude Fable chatbot, the Associated Press quoted him saying.
Why Americans choose Chinese artificial intelligence
Before switching to Kimi K3, Krikorian had been using another Chinese AI model, Z.ai’s GLM-5.2, for daily tasks such as managing calendars, documents and emails. Krikorian is not alone. Now more and more Americans are adopting Chinese AI systems because they offer good performance at a much lower price. U.S. cryptocurrency exchange Coinbase has also begun using Chinese artificial intelligence models to reduce operating costs.
China’s artificial intelligence models are increasingly attractive to software developers because they are affordable, even as some U.S. tech companies chafe at their growing popularity. Unless the U.S. bans these AI models entirely, they are likely to remain popular with developers in the U.S. and other countries, the report said. The United States has blocked China from purchasing some of the world’s most advanced artificial intelligence chips and other key technologies.
Trump clamps down on Chinese artificial intelligence
U.S. Treasury Secretary Scott Bessant warned that the Trump administration may impose more sanctions to protect U.S. intellectual property.
On Wednesday, the Trump administration accused moon landing According to the Associated Press, a “secret” method was used to build the Kimi K3 using Anthropic’s Fable model. The government says these methods are not necessarily illegal. Several U.S. politicians and AI companies, including Anthropic, have accused Chinese AI startups of using “distilled” technology to copy U.S. AI models.
China has denied the accusations, calling them “baseless.” The competition between China and the United States in the field of artificial intelligence has become more intense since Chinese startup DeepSeek surprised the tech world earlier this year. DeepSeek has gained attention for offering an artificial intelligence model that performs similarly to top U.S. models but at a much lower cost. Chinese companies have since launched multiple advanced AI models, including Z.ai’s GLM-5.2 in June and Moonshot’s Kimi K3 in July. Alibaba also launched a preview version of the Qwen3.8 Max AI model in July.
DeepSeek also released a preview of its latest V4 model in April to work with OpenAI’s GPT and Google Gemini. Experts told The Associated Press that these latest Chinese AI models are now nearly as powerful as the most advanced AI systems built by OpenAI and Anthropic.
Curt Meinhold, founder and technical director of LilyList from North Carolina, said he now generally prefers to use DeepSeek to perform business tasks such as finding sales leads. Meinhold believes that most people don’t need the most powerful and expensive AI models. As quoted by the AP, he said: “The bottom line is that most of us, the vast majority of us, over 90 percent, don’t need (human) myth or fable.” He added, “Like, we just don’t need it, we need something good enough.”
Chinese artificial intelligence is booming around the world
Investment bank Goldman Sachs said in a July research note that China’s artificial intelligence models are at a “critical stage” for wider global adoption. Goldman Sachs said demand for cheaper AI models is increasing as more companies use AI agents that automate complex tasks. Analysts such as Alex Colville of the Australian Strategic Policy Institute say AI pricing is based on the number of tokens processed, making cheaper models more attractive to businesses.
The Associated Press quoted Meinhold as saying, “You know, I find the Chinese model to be very close in terms of code and research.” He added, “If I can pay a few cents per million output tokens instead of $30, $40 or $50, that’s enough.”
Data from artificial intelligence platform OpenRouter shows that the five most popular artificial intelligence models in the past month are all Chinese. Market intelligence firm Sensor Tower estimates that Kimi was downloaded more than 930,000 times within a week of K3’s launch. This is a 200% increase compared to the previous week.
In the United States alone, “Kimi” has been downloaded approximately 86,000 times, a 387% increase from the previous week. Kimi K3 became so popular that Moonshot temporarily stopped accepting new subscriptions due to high demand. The Associated Press quoted Arena CEO Anastasios Angelopoulos as saying that despite the rapid progress, experts say China’s artificial intelligence models still lag behind the best American artificial intelligence systems in terms of overall capabilities.
Can Chinese AI beat ChatGPT?
The Associated Press quoted Yasser Attaran from the Center for Strategic and International Studies as saying that experts also said that American artificial intelligence companies are developing cheaper versions of their own models to stay competitive. One of China’s great advantages is that most Chinese AI models are open source and anyone can inspect, modify and build them. By contrast, leading U.S. AI models from companies like OpenAI and Anthropic remain closed source.
Technology research firm Omdia expects Chinese companies to use open source artificial intelligence to expand their global influence. The Associated Press quoted Krikorian as saying that “open borders are increasingly being built by China.”
Despite U.S. debate over restrictions on Chinese AI models, including MicrosoftMeta and Nvidia signed an open letter supporting open AI models. The report stated that some U.S. government policies may inadvertently help Chinese artificial intelligence companies gain more users. For example, Anthropic’s Fable and Mythos models were unavailable for more than two weeks after the Trump administration imposed export controls on them, while China’s Z.ai launched its GLM-5.2 model during that period.
Arena CEO Anastasios Angelopoulos said “limiting the U.S. model could immediately create opportunities for Chinese competitors,” according to the Associated Press. Within China, businesses and consumers have rapidly adopted AI products with strong government support. Chinese technology companies such as Huawei and Tencent are adding artificial intelligence capabilities to smartphones, AI glasses and humanoid robots.
China promotes global artificial intelligence development
China is also trying to expand its AI influence globally. At a major technology summit in Shanghai, Chinese President Xi Jinping promoted open source artificial intelligence and said China wanted to help developing countries improve their artificial intelligence capabilities. Fierce competition within China is driving artificial intelligence startups to expand into overseas markets. Many leading Chinese artificial intelligence companies are raising new capital, including through public stock offerings, to support their global expansion.
Chelsea Tan of Morningstar said both the United States and China want more countries to adopt their artificial intelligence ecosystems while protecting technology that can help strategic rivals. Angelopoulos said the competition is no longer just between the United States and China, as Chinese AI companies are also competing fiercely with each other.
Some experts also warn that huge spending on AI raises questions about the long-term financial sustainability of Chinese AI startups. Zaiai Technology, for example, reported that its revenue grew 132% last year to 724 million yuan ($107 million), but its net loss also increased 60% to 4.7 billion yuan ($694 million).
Despite political tensions, trade restrictions and concerns about technology theft, experts believe China’s AI models are likely to continue to win users in the United States because they offer strong performance at much lower prices. Krikorian encouraged others to try China’s AI models, saying: “I strongly recommend that anyone who is serious about doing AI payloads at least evaluates it,” Krikorian was quoted as saying by the Associated Press.
