US Federal Agencies Ordered to Remove DeepSeek AI Systems
National security concerns drive federal mandate to purge Chinese AI technology from government infrastructure.
1Federal Ban on DeepSeek
The U.S. Department of Homeland Security has officially ordered all federal civilian agencies to cease purchasing and utilizing products developed by the Chinese startup DeepSeek. This directive mandates the removal of DeepSeek technology from existing government IT systems, citing significant national security concerns regarding the company's operations and potential data handling practices.
2Scope of the Restriction
The prohibition extends beyond direct DeepSeek products to include any third-party services or software built upon their AI technology. While federal agencies are strictly required to comply with these new security protocols, the order does provide a formal mechanism for agencies to request specific exceptions if operational requirements necessitate continued use of tools.
3Data Privacy and Risks
The primary justification for this mandate involves the potential exposure of sensitive government information. Officials warned that user data processed through DeepSeek systems could become accessible to Chinese authorities under local legal frameworks. This risk assessment triggered the expansion of restrictions from intelligence-specific systems to the broader federal civilian agency infrastructure.
4Escalating Tech Competition
This policy move reflects the intensifying AI rivalry between the United States and China. DeepSeek has emerged as a significant competitor to American firms like OpenAI and Meta by offering low-cost, open-weight models. Beijing has formally opposed the U.S. decision, characterizing it as an overextension of national security concepts into trade.
5Historical Regulatory Context
Restrictions on DeepSeek have evolved rapidly since a December 2025 law initially targeted the startup within national security and intelligence systems. The current directive represents a major escalation, covering all civilian federal operations. This follows a September report accusing DeepSeek and five other Chinese firms of extracting data from American AI systems.
6Model Distillation Allegations
Major U.S. AI developers, including OpenAI and Anthropic, have raised alarms regarding DeepSeek's alleged use of model distillation. This technique involves using outputs from established models to train competing systems. OpenAI reported observing suspected distillation attempts, while Anthropic identified unauthorized efforts by Chinese labs to collect outputs from their Claude model.
7Chinese Export Countermeasures
In response to tightening U.S. controls, China is reportedly considering its own export restrictions on 23 AI-related technologies. These potential measures include controls on chip designs and advanced AI models. This reciprocal strategy aims to address internal security risks and prevent the misuse of domestic technological advancements by foreign entities and competitors.
8Global Industry Impact
DeepSeek has faced intense international scrutiny since its R1 reasoning model gained prominence in January 2025. The model's rapid adoption triggered a notable sell-off in Nvidia shares, highlighting the market's sensitivity to Chinese AI progress. Governments worldwide are now re-evaluating their reliance on these models amid ongoing concerns regarding intellectual property and security.
Original article: TechCrunch — US Federal Agencies Ordered to Remove DeepSeek AI Systems