US Bans DeepSeek From Civilian Government Systems
The Department of Homeland Security is banning civilian federal agencies from using or buying products and services from DeepSeek, including third-party services built with its AI models, Reuters reports. Agencies must also remove DeepSeek products already in use, and the company is barred from federal contracts, according to Engadget.
The notice cites the Federal Acquisition Supply Chain Security Act and a review by the Federal Acquisition Security Council, and is set to publish in the Federal Register on Tuesday, CyberScoop reports. It includes a letter from Homeland Security Secretary Markwayne Mullin dated June 26 ordering the exclusion, so the ban has applied for months. Agencies can request exceptions 60 days in advance with a "compelling justification".
DHS said DeepSeek's models pose "multiple threat vectors": data is stored on servers in China, where the government can demand access, and use by agencies "may result in the transfer of sensitive data" to the Chinese government. It added that the models "may contain malware or other security vulnerabilities." The order covers civilian agencies such as Treasury, State and Justice; Defense agencies are excluded, according to BBC News.
Several agencies, including the Commerce Department and the Navy, had already banned DeepSeek on government devices, as had states including New York, Tennessee and Virginia. The order follows criticism of Chinese AI firms by the Trump administration, and comes days after DeepSeek released its V4.1-Flash model.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like AI Policy Brief, Cybersecurity AI Weekly or Daily AI Brief.
Also, consider following us on social media:
More from Regulation
Oct 19 NIST, NSA and CISA Issue Super Intelligence Incident Guidance Oct 19 Canadian News Publishers Sue Anthropic Over Claude Training Oct 19 Vanderbilt Report Proposes Defense Production Act Controls for AI Chip Demand Oct 19 Denmark Plans Broad AI Data Powers for Public Bodies Oct 19 Wisconsin DNR to Review AI Data Center Impacts StatewideMore from Cybersecurity
Oct 19 NIST, NSA and CISA Issue Super Intelligence Incident Guidance Oct 19 Darktrace Survey Finds Formal AI Policies Declined to 37% Oct 19 Claude Code Detects Crypto Miner Using About 12 Mac CPU Cores Oct 19 AI Search Tools Cited a Fictional Concept Within One Week Oct 19 SandboxAQ Adds Agent Security and ROI Tools to Switch AIAI Policy Brief
Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.
Whitepaper
Tensordyne Napier: What If One Rack Could Do the Work of Nine?
Tensordyne
This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.
Read moreYou may also like
US Allows Britain to Test American AI Models Before Release
DeepSeek Releases V4.1 Flash 1017 for Coding and Agent Tasks
White House mandates AI incident reporting after Anthropic model incidents
Chinese AI Agents Deceived Evaluators in Controlled Tests
OpenAI apologizes for access to Australian government systems
Daily AI Brief: the AI news that matters, in your inbox.