DHS Orders Federal Agencies To Drop DeepSeek AI

The Department of Homeland Security ordered federal agencies to stop buying DeepSeek AI products and strip them, including third-party tools built on DeepSeek technology, from government systems over national-security concerns.

DHS Orders Federal Agencies To Drop DeepSeek AI: key figures
ShareXLinkedInEmail

THE RUNDOWN

1

The order covers third-party products built with DeepSeek technology, not just DeepSeek's own apps, which could sweep in U.S. software vendors who quietly used DeepSeek models as a backend without disclosing it.

2

A December 2025 law already barred DeepSeek from intelligence and national-security systems, so this order extends the same logic to the much larger population of ordinary federal civilian agencies.

3

Beijing's response -- accusing Washington of 'politicizing technology and trade issues' -- signals the DeepSeek dispute is now layered into the broader US-China AI and chip-export standoff rather than a standalone security review.

4

Anthropic and OpenAI have both said they identified attempts by DeepSeek and other Chinese labs to distill outputs from their models, so this order lands alongside a parallel industry fight over model distillation, not in isolation.

The VC Read

Value Add VC analysis

The enforcement gap is the real story: DHS banned DeepSeek-built products without saying how agencies verify a vendor isn't quietly running DeepSeek under the hood. Any founder selling AI tooling into federal agencies should be checking their own model supply chain now, not waiting for an audit to find it for them.

Analysis

The Department of Homeland Security issued an order, published in a Federal Register notice dated October 20, 2026, directing federal civilian agencies to stop purchasing DeepSeek products and services and to remove them from government IT systems, according to TechCrunch. The order reaches further than DeepSeek's own apps: it also covers third-party products and services built using DeepSeek technology, applies to both new purchases and systems already in place, and allows agencies to request exceptions.

DHS's stated rationale is that sensitive data handled by U.S. government users could potentially be accessible to Chinese authorities under Chinese law -- the same national-security logic that has driven scrutiny of DeepSeek across multiple countries since its R1 model launched in January 2025 and triggered a sharp sell-off in Nvidia shares. A December 2025 law had already restricted DeepSeek's use in intelligence and national-security systems specifically; this order extends similar restrictions to the much broader set of ordinary federal civilian agencies that law didn't cover.

“DHS's stated rationale is that sensitive data handled by U.S.”

This is one front in a wider US-China AI fight, not an isolated ban

China's Foreign Ministry pushed back through spokesperson Mao Ning, saying Beijing opposes "overstretching the concept of national security and politicizing technology and trade issues" and that it would defend Chinese enterprises' rights. The exchange lands alongside reports that China is considering export controls on 23 AI-related technologies, including chip designs and certain advanced models, and alongside a September report tied to a CISA advisory accusing DeepSeek and five other Chinese AI firms -- including Alibaba and Moonshot AI -- of large-scale model distillation aimed at extracting information from US AI systems. Both OpenAI and Anthropic have separately said they identified suspected DeepSeek attempts to distill outputs from their own models.

Pulse has covered DeepSeek's push into open-weight models and pre-IPO fundraising as the company keeps shipping product even as Western governments tighten restrictions around it -- a split-screen that defines the current DeepSeek story: commercially competitive model releases on one side, expanding bans on the other. The order doesn't name a compliance deadline, and DHS hasn't said how agencies are expected to verify that a given software product doesn't secretly rely on DeepSeek technology somewhere in its stack -- an enforcement question that may prove harder than writing the policy itself.

ShareXLinkedInEmail

Key Sources

1 source

Reported by TechCrunch · Analysis by Value Add Pulse.

← Back to Pulse

THE WIRE in your inbox— Tech, startup & VC news with The VC Read, a few times a week. Free to subscribe, no spam.

No spam, unsubscribe anytime. Sent via Substack · Privacy · or subscribe on Substack ↗