US companies are increasingly adopting Chinese AI models for their operations, with firms like Pinterest and Airbnb integrating these technologies into their systems.
Adoption by Major Companies
Pinterest has been experimenting with Chinese AI models, including DeepSeek R-1, to enhance its recommendation engine.
According to Pinterest CTO Matt Madrigal, open-source techniques derived from these models offer 30% greater accuracy and can be up to 90% more cost-effective than proprietary models from US developers such as OpenAI.
The availability of open-source models like DeepSeek R-1, which launched in January 2025, has contributed to a wider trend of accessible AI technologies.
Airbnb's AI Strategy
Airbnb CEO Brian Chesky confirmed in October that his company utilizes Alibaba's Qwen model for its AI customer service agent, noting its effectiveness, speed, and affordability.
Airbnb uses multiple AI models, including US-based ones, hosting them securely within its infrastructure and ensuring data is not shared with AI model developers.
Market Presence and Cost Advantages
Chinese AI models are gaining significant traction, particularly due to their cost-effectiveness.
On Hugging Face, a prominent platform for downloading AI models, Chinese models from labs like Alibaba frequently appear among the top-ranked and most downloaded options.
Jeff Boudier from Hugging Face highlighted that cost is a primary factor driving startups to choose Chinese models over their US counterparts.
In September, Alibaba's Qwen family of large language models surpassed Meta's Llama in download popularity on the platform.
Meta's Llama models, open-sourced in 2023, were previously a leading choice for developers. However, the release of Llama 4 reportedly generated less enthusiasm among developers. Meta is reportedly training a new model by incorporating various open-source models, including those from Alibaba, Google, and OpenAI.