US orders federal agencies to remove China’s DeepSeek AI from their systems

The U.S. has ordered national civilian agencies to halt purchasing services and products developed by Chinese AI startup DeepSeek and region them from authorities IT systems implicit nationalist information concerns tied to the Chinese AI startup’s operations successful China, according to the Department of Homeland Security. The bid goes beyond banning DeepSeek’s ain product, but also covers third-party products and services built utilizing DeepSeek technology, though agencies tin petition exceptions.

The DHS raised concerns astir DeepSeek’s handling of idiosyncratic data, informing that delicate accusation from US authorities users could perchance beryllium accessible to Chinese authorities nether Chinese law, the bid said.

The determination comes amid intensifying tech contention successful AI betwixt the US and China. The Chinese AI exemplary challenges its U.S. rivals, including OpenAI and Meta, with low-cost open-weight models.

China pushed backmost against the U.S.’s move. Foreign Ministry spokesperson Mao Ning said Tuesday that Beijing opposes “overstretching the conception of nationalist information and politicizing exertion and commercialized issues,” according to an authoritative briefing. “We volition resolutely safeguard the morganatic rights and interests of Chinese enterprises,” the spokesperson added.  

The U.S. has already taken steps to bounds DeepSeek’s usage successful government. A instrumentality passed successful December 2025 targeted its usage successful intelligence’s nationalist information systems. The latest bid extends the restrictions to national civilian agencies, covering some caller purchases and existing IT systems.

Meanwhile, China is reportedly considering export restrictions connected 23 AI-related technologies, including spot designs and definite precocious AI models, to code information and misuse risks, Bloomberg reported past week.

The Chinese AI laboratory has faced scrutiny from governments and regulators crossed the globe since its R1 reasoning exemplary shook the AI manufacture in January 2025, triggering a crisp sell-off successful Nvidia shares.

The U.S. accused DeepSeek and 5 different Chinese AI companies, including Alibaba and Moonshot AI, of utilizing large-scale exemplary distillation to extract accusation from American AI systems, according to a September report.

In addition, OpenAI and Anthropic person separately raised concerns astir DeepSeek’s alleged usage of exemplary distillation, a method that uses outputs from 1 AI exemplary to bid another. Earlier this year, OpenAI told U.S. lawmakers it had observed suspected attempts by DeepSeek to distill its models, portion Anthropic said it had identified unauthorized attempts by DeepSeek and different Chinese AI labs to cod outputs from Claude.