Chinese AI Models are edging out U.S. Rivals in Silicon Valley
Chinese artificial intelligence systems DeepSeek, Qwen, and Kimi are rapidly displacing American competitors from Silicon Valley.
More and more Western companies are choosing Chinese models because of their affordability, speed, and flexibility in customising them for their own tasks.
Despite the ongoing tech rivalry between the United States and China, U.S. businesses are increasingly turning to China-made solutions.
Airbnb CEO Brian Chesky admitted in October that the company is using Alibaba's Qwen model line to create its own in-app assistant. According to him, these models are “very fast, high-quality, and cheap.”
At the same time, he stressed that integration with ChatGPT does not yet meet Airbnb's business needs.
The investment fund Social Capital, led by Chamath Palihapitiya, took a similar position. The organisation transferred part of its computing processes to the Chinese model Kimi K2. Palihapitiya explained the decision by saying that it is “significantly more productive and many times cheaper” than OpenAI and Anthropic products.
Data from the Hugging Face analytics platform, prepared for the ATOM Project initiative, confirms the scale of the trend. While Meta's Llama model had about 10.6 million downloads at the beginning of 2024 and Qwen had only half a million, by the autumn of 2025 the situation had completely changed.
Qwen downloads reached 385.3 million, while Llama received 346.2 million. Moreover, models based on Qwen already account for more than 40% of new language solutions on the Hugging Face platform, while the share of the Llama family is limited to about 15%.
If you want to be completely digitally secure when communicating with family or friends, download the app at the following links: for iOS; for Android; for Windows.
As a reminder, Alibaba's Qwen garnered over 10 million downloads in a week.
OpenAI also introduced its own browser with artificial intelligence.
