The U.S. government has ordered federal civilian agencies to stop purchasing services and products from Chinese AI startup DeepSeek and remove them from government IT systems, according to the Department of Homeland Security (TechCrunch). The directive addresses national security concerns tied to the company's operations in China and user data handling.
The order extends beyond DeepSeek's direct offerings to cover third-party products and services built using its technology, though agencies can apply for exceptions. Officials raised concerns that sensitive information from U.S. government users could potentially be accessible to Chinese authorities under local laws.
This latest action builds on a December 2025 law that targeted DeepSeek's use within intelligence and national security systems. The new directive expands those restrictions across federal civilian executive branch information technology systems and procurements.
Tensions have mounted amid broader allegations regarding model distillation. U.S. security agencies previously stated that DeepSeek and other Chinese firms conducted industrial-scale knowledge distillation campaigns against American AI models (Cybersecurity and Infrastructure Security Agency CISA).
Anthropic reported that DeepSeek, Moonshot, and MiniMax generated over 16 million exchanges with Claude through thousands of fraudulent accounts to illicitly extract capabilities (Anthropic). Anthropic noted that while distillation is a legitimate training method, illicit use can strip away important safety safeguards.
China pushed back strongly against the U.S. restrictions. Foreign Ministry spokesperson Mao Ning stated Tuesday that Beijing opposes overstretching national security concepts to politicize technology and trade, adding, "We will resolutely safeguard the legitimate rights and interests of Chinese enterprises".
The ongoing dispute highlights intensifying technological competition between the U.S. and China, particularly surrounding low-cost open-weight models. Meanwhile, reports indicate China is also considering export restrictions on various advanced AI technologies and chip designs.
The federal directive against DeepSeek marks a significant escalation in the regulatory containment of foreign AI developers. By extending restrictions from national security systems to all civilian agencies, the U.S. government is signaling that data exposure risks and industrial-scale distillation campaigns are viewed as systemic threats. This policy alignment between trade controls and cybersecurity measures suggests that integration of foreign open-weight models in enterprise settings will face mounting compliance hurdles.
Spotted an error? See our corrections policy or email editorial@lordofclaude.com.
Discussion & Comments (0)
Join the conversation