Chinese Open-Weight AI Models Gaining US Enterprise Adoption
Key Questions
Why are US companies adopting Chinese open-weight AI models?
US enterprises are shifting to models like DeepSeek, Z.ai's GLM-5.2, Moonshot, and MiniMax due to substantial cost savings compared to OpenAI and Anthropic offerings. This aligns with broader industry moves toward cheaper open-source alternatives, as noted by Amazon's CTO.
What market share have Chinese AI models achieved in US enterprise usage?
Chinese-origin models now capture up to 46% of US enterprise token usage on OpenRouter. They have accounted for at least 30% of such usage every week since February 8, 2026.
How does this trend relate to access to affordable AI models?
The adoption reflects growing interest in underrated models that provide free or low-cost access. It supports efforts to reduce mounting AI expenses through open-weight alternatives.
A wave of Chinese models (DeepSeek, Z.ai's GLM-5.2, Moonshot, MiniMax) are being adopted by US companies due to massive cost savings over OpenAI/Anthropic. New data shows Chinese models now capture up to 46% of US enterprise token usage on OpenRouter. This trend aligns with the curator's focus on underrated models with free/cheap access.