US firms increasingly adopt Chinese AI models as security concerns ease

The hesitation among US businesses over Chinese AI offerings appears to be diminishing. Harvey, a legal-AI startup, announced a new internal model built partly on Moonshot’s Kimi K3 technology. At the same time, Snowflake expanded its service to include models from DeepSeek and Z.ai, enabling automatic selection of the most cost-effective or highest-performing engine for each job.

Snowflake’s head of security and trust, Mayank Upadhyay, told reporters in New York that the company’s hosting environment and built-in safeguards lessen concerns about data privacy or Beijing-centric influence. Fanatics’ vice president of data and AI emphasized that the low price of Chinese models is a major attraction and that companies need nuanced policies rather than outright bans. These moves signal growing confidence in the practicality and safety of foreign AI services.