Source: Fortune
Summary
Companies like DoorDash, Airbnb, and Siemens are turning to cheaper, open-source AI models from Chinese companies like Moonshot AI and DeepSeek to save on rising AI costs.
These models offer better quality and lower costs, but also raise concerns about security risks and data sovereignty.
Experts warn that adopting these models may expose proprietary code and user data to foreign surveillance.
Despite this, companies are experimenting with alternative models for different tasks, and some are finding that running open-source models locally can give them more control over their data.
Our Reading
The numbers tell one story.
DoorDash, Airbnb, and Siemens are joining the likes of Cursor and Lindy in adopting cheaper, open-source AI models from Chinese companies.
Moonshot AI’s Kimi and DeepSeek’s V4 models are gaining traction, with 41% of downloads on Hugging Face coming from Chinese open-source models.
But experts like Snehal Antani warn of severe data sovereignty violations and critical vulnerabilities in model integrity and reasoning.
Yasir Atalan notes that companies are experimenting with alternative models for different tasks, rather than replacing US models outright.
The trend is not a wholesale migration to Chinese AI models, but rather a search for cost-effective solutions.
The strategy enters a familiar phase: cost-cutting in the name of innovation, with security risks and data sovereignty concerns relegated to the fine print.
Author: Evan Null
