The article reports that companies are increasingly adopting open-weight models developed in China as a strategy to reduce operational expenses.
This trend highlights a shift in the industry where cost-efficiency drives the selection of AI infrastructure, specifically favoring accessible model weights over proprietary alternatives.