Lead (5W1H) On Tuesday, the U.S. Department of Homeland Security issued a directive that requires all federal civilian agencies to cease purchasing any services or products developed by Chinese artificial‑intelligence startup DeepSeek and to remove existing DeepSeek‑based tools from their IT environments. The order, motivated by concerns that user data could be accessed by Chinese authorities under Chinese law, applies not only to DeepSeek’s own offerings but also to third‑party solutions that incorporate its technology, though agencies may request exceptions.
Scope of the DHS Directive
According to TechCrunch, the Department of Homeland Security’s order explicitly bans the acquisition of DeepSeek‑originated AI products across civilian federal agencies and mandates the deletion of those tools from current government systems. The prohibition extends to any third‑party software that leverages DeepSeek’s underlying models, underscoring the breadth of the security rationale. Agencies are permitted to file for exemptions on a case‑by‑case basis, but the default position is a complete removal.National‑Security Rationale
The DHS briefing, as reported by TechCrunch, highlighted the risk that “sensitive information from US government users could potentially be accessible to Chinese authorities under Chinese law.” This concern centers on DeepSeek’s handling of user data while operating under Chinese jurisdiction, where legal frameworks may compel companies to share information with the state. The agency’s assessment links these data‑privacy issues directly to national‑security considerations.Context of the U.S.–China AI Competition
TechCrunch notes that the directive arrives amid an intensifying technology rivalry between the United States and China in the field of artificial intelligence. DeepSeek’s models, described as low‑cost and open‑weight, have been positioned as challengers to U.S. incumbents such as OpenAI and Meta. The competitive pressure, combined with security worries, has prompted Washington to tighten controls on Chinese AI imports.Beijing’s Reaction
In response, China’s foreign ministry spokesperson Mao Ning told reporters, as cited by TechCrunch, that Beijing “opposes overstretching the concept of national security and politicizing technology and trade issues.” The spokesperson added that China will “resolutely safeguard the legitimate rights and interests of Chinese enterprises,” signalling diplomatic pushback against the U.S. measures.Legislative Background
The current order builds on earlier legislative action. According to TechCrunch, a law enacted in December 2025 already barred the use of DeepSeek technology in U.S. intelligence and national‑security systems. The new directive expands that prohibition to the broader civilian federal apparatus, covering both fresh procurements and existing installations.Parallel Moves in China
TechCrunch also reported that China is reportedly weighing export restrictions on 23 AI‑related technologies, including certain chip designs and advanced AI models. The move, described as a response to security and misuse concerns, was attributed to Bloomberg in the original source. While the specifics of China’s export plan remain under discussion, the report suggests a reciprocal tightening of AI‑related trade controls.Industry Scrutiny of DeepSeek
Since the launch of DeepSeek’s R1 reasoning model in January 2025, the company has attracted global regulatory attention, according to TechCrunch. The model’s performance sparked a sharp sell‑off in Nvidia shares, illustrating the market impact of the Chinese AI lab’s rapid progress. Moreover, the U.S. has accused DeepSeek and five other Chinese AI firms—including Alibaba and Moonshot AI—of employing large‑scale model distillation to extract information from American AI systems, as detailed in a September report referenced by TechCrunch.Concerns from U.S. AI Leaders
OpenAI and Anthropic have each voiced apprehensions about DeepSeek’s alleged use of model distillation, a technique that repurposes outputs from one model to train another. TechCrunch notes that earlier this year OpenAI briefed U.S. lawmakers on suspected attempts by DeepSeek to distill its models, while Anthropic reported unauthorized collection of outputs from its Claude system by DeepSeek and other Chinese labs. These industry‑level warnings reinforce the government’s security narrative.--- Opinion (Analysis) The DHS directive reflects a broader strategic shift in U.S. policy that treats advanced AI technologies not merely as commercial products but as potential vectors for espionage. By extending restrictions from intelligence agencies to civilian departments, Washington signals that data sovereignty concerns now permeate all levels of federal operation. While the move may safeguard sensitive information, it also risks fragmenting the global AI ecosystem and could accelerate a bifurcation of AI standards between the West and China. The diplomatic backlash from Beijing underscores the delicate balance between security imperatives and international trade norms.
--- Conclusion The Department of Homeland Security’s order to purge DeepSeek AI from federal civilian agencies marks the latest escalation in the U.S.–China technology rivalry. Citing data‑privacy risks under Chinese law, the directive expands earlier restrictions and invites exceptions only on a limited basis. Beijing’s verbal protest and parallel considerations of export controls suggest a tit‑for‑tat environment that may reshape the global AI landscape in the months ahead.
Source: TechCrunch
Frequently Asked Questions
Which U.S. agencies are affected by the new DeepSeek restriction?**
All federal civilian agencies are required to stop purchasing DeepSeek products and to remove existing DeepSeek‑based tools from their IT systems, according to TechCrunch. **Q: Can agencies still use DeepSeek technology if they obtain an exception?** A: Yes, the DHS order allows agencies to request exemptions on a case‑by‑case basis, as reported by TechCrunch. **Q: What specific security concern did the DHS cite regarding DeepSeek?** A: The DHS warned that user data processed by DeepSeek could be accessible to Chinese authorities under Chinese law, posing a national‑security risk, according to TechCrunch.