On February 27, 2026, Secretary of War Pete Hegseth said he was directing the Department of War to designate Anthropic a national-security supply-chain risk after talks over Claude’s military use broke down. Anthropic said it had not received formal notice of a completed designation. The dispute centered on the company’s refusal to remove restrictions on mass domestic surveillance and fully autonomous weapons, not on military use of AI generally.
What happened—and what was not yet confirmed
Anthropic’s February 27 statement said Hegseth had directed the Department of War to make the designation. The company also said it had received no direct communication from the department or the White House about the status of negotiations and would challenge a formal designation in court. The distinction matters: a public direction to designate is not, by itself, proof that a final statutory determination or contracting instruction took effect. Anthropic’s statement and contemporaneous reporting described the announced action, but the available accounts do not establish the text or completion of a formal designation, a written contractor prohibition, or a court filing.
| Date | Event | Status and attribution |
|---|---|---|
| June 2024 | Anthropic says it began supporting U.S. warfighters and deployed models in classified government networks. | Anthropic’s account; it does not establish authorization for unrestricted uses. |
| February 27, 2026 | Anthropic said negotiations had reached an impasse and that Hegseth directed the department to designate it a supply-chain risk. | Publicly announced direction; Anthropic said formal direct notice had not arrived. |
| February 28, 2026 | OpenAI announced an agreement with the Department of War for advanced AI deployment in classified environments. | OpenAI’s description of its own agreement, not an independent audit. |
| March 2, 2026 | OpenAI said it updated its agreement to make the domestic-surveillance restriction more explicit. | OpenAI’s account. |
Separate reports described a White House plan for agencies to phase out Anthropic technology over six months and broader messaging to contractors about commercial activity. Those reported measures should not be conflated with the Department of War designation: the available accounts do not establish the underlying written directive or its precise scope. The reported chronology is not a substitute for the actual procurement notice or memorandum.
Why negotiations broke down
Anthropic said it supported lawful national-security work, including foreign-intelligence and counterintelligence uses, but would not authorize two categories: mass domestic surveillance of Americans and fully autonomous weapons. The Pentagon’s stated position, as reported, was that it needed access for all lawful purposes and that a private model provider should not impose restrictions that could interfere with military operations. Spokesperson Sean Parnell reportedly said the department did not seek mass domestic surveillance or autonomous weapons without human involvement. That account should be read as the department’s stated position, not as resolution of whether contract safeguards were enforceable or what future policy might permit.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
- Compact and Efficient Design: The FortiGate 40F is designed for small to mid-sized businesses and enterprise branch offices, featuring a compact, fanless desktop form factor that ensures quiet operation and minimizes space usage.
- Robust Connectivity Options: Equipped with 5 GE RJ45 ports, including 1 WAN port and 4 internal ports, this model provides essential connectivity and flexibility for various network configurations in a small-scale environment.
- High-Performance Security: Offers up to 1 Gbps IPS throughput and 600 Mbps threat protection throughput, using Fortinet’s purpose-built security processor technology to deliver industry-leading performance and protection for SSL encrypted traffic.
- Advanced Threat Protection: Integrated with Fortinet’s AI-powered FortiGuard Labs, the FortiGate 40F offers comprehensive cybersecurity, identifying and mitigating both known and unknown threats to maintain robust security across your network.
- Simplified Management and Deployment: Features a user-friendly management console that provides comprehensive network automation and visibility, coupled with Zero Touch Integration with Fortinet’s Security Fabric for easy deployment.
Mass domestic surveillance
Anthropic drew a line between targeted, lawful intelligence activity with a defined foreign-intelligence purpose and broad monitoring or analysis of Americans’ private or commercially acquired personal information. Its objection was to the latter category, which it said posed serious civil-liberties risks. That position is not opposition to all intelligence analysis or all government use of Claude. Anthropic’s explanation sets out its distinction.
Fully autonomous weapons
Anthropic argued that current frontier models are not reliable enough to make lethal decisions without meaningful human responsibility and control. “AI in military systems” is much broader than a system independently selecting and engaging targets: AI might support logistics, software development, intelligence analysis, or other work without controlling a weapon. The disputed boundary concerned removing meaningful human control and accountability from the use of force. Anthropic’s position and rationale are described in its February 27 statement.
What a supply-chain-risk designation could mean
Anthropic cited 10 U.S.C. § 3252, a statute concerning national-security risks involving products, services, and suppliers in the defense supply chain. The legal and operational consequences depend on the authority invoked and the terms of any final determination or implementing instruction; the announcement alone does not establish a universal commercial blacklist. The current statutory text is available from the U.S. Code, and agency documentation would be needed to establish how the authority was applied in this case.
Rank #2
- HARDWARE PLUS SECURITY SERVICES: FortiGate-60F Firewall Appliance bundled with 1 year of FortiCare Premium and FortiGuard Unified Threat Protection.
- UNIFIED THREAT PROTECTION (UTP): Secures against advanced online threats with comprehensive web filtering and anti-botnet technologies.
- OPTIMIZED FOR MEDIUM-SIZED BUSINESSES: Tailored for businesses needing robust security without the infrastructure of larger enterprises.
- RELIABLE CUSTOMER SUPPORT: FortiCare Premium ensures high-quality support and service continuity.
- EFFECTIVE PROTECTION: Employs advanced filtering technologies to safeguard against sophisticated threats.
Anthropic’s interpretation was that a designation would limit Claude’s use in Department of War contract work, not prohibit it from serving unrelated customers. It said a contractor could continue using Claude for work unrelated to the department. That is the company’s legal and operational view, not a court ruling or a confirmed government-wide instruction. Anthropic’s statement explains its reading of the statute.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minute| Customer or use | Potential issue | What is established |
|---|---|---|
| Department of War | Procurement, deployment, classified-network access, and Claude use on covered contracts could be affected. | The department’s final written terms and scope are not established by the public announcement alone. |
| Prime contractors and subcontractors | They might need to assess whether Claude is used in covered work, delivered systems, internal tools, or contract performance. | A blanket prohibition across all commercial activity was not established in the available accounts. |
| Other commercial customers | Continuity and contractual assurances may matter if a customer also does defense work. | Anthropic said ordinary commercial customers and unrelated work would be unaffected; that remains its interpretation. |
| Other federal agencies | A reported six-month phase-out could affect separate agency deployments. | The underlying written directive and its scope were not established in the available accounts. |
| Cloud and software vendors | Providers may need to determine whether hosting or embedding Claude touches covered contract work. | No specific rule for these arrangements was established in the available accounts. |
For a contractor, the practical question is not simply whether anyone at the company uses Claude. It is whether a particular use supports covered Department of War work, is incorporated into a deliverable, or falls under a contract certification or instruction. A model used for internal software development, a commercial platform with government customers, and a Claude deployment on a classified network raise different questions. Until the governing directive and contract language are available, organizations should not assume either that all use is barred or that separating workloads automatically resolves compliance.
The competing arguments over control and risk
The Pentagon’s stated case
The department’s position, as reported, was that military users need access for all lawful missions and that provider-imposed limits could constrain operational readiness. Under that view, government officials—not a vendor—should decide which lawful missions can use a model. A further concern is dependence on a supplier whose terms might change or prevent use in a time-sensitive operation.
Rank #3
- 【Up to 1100 Mbps VPN Speed 】 Hardware-accelerated WireGuard and OpenVPN-DCO deliver up to 1100 Mbps VPN throughput, over 3× faster than Brume 2 for smooth remote access and file transfers.
- 【Three 2.5G Ports & Multi-WAN】Tri-port 2.5GbE design with flexible WAN LAN configuration supports multi-gigabit wired setups, dual-ISP Multi-WAN and failover to keep home and SOHO networks online.
- 【Stealth VPN Obfuscation】VPN obfuscation disguises VPN traffic as regular HTTPS, helping you evade blocking, bypass restrictive networks and maintain stable, private connections.
- 【DPI protection】Deep Packet Inspection with visual dashboards blocks adult/gambling/malicious sites, while SQM and QoS prioritize gaming, calls, and video when bandwidth is tight
- 【OpenWrt & USB 3.0 Expansion】OpenWrt with 1GB DDR4 and 8GB eMMC lets you install plugins and build VPN, ad-blocking or NAS, while USB 3.0 Type‑C connects high-speed storage or 4G/5G dongles
Anthropic’s case
Anthropic argued that legal permission is not the same as technical reliability or acceptable risk. It said safeguards should preserve human responsibility for lethal force and protect Americans from mass surveillance. The company also warned that using supply-chain authority against a U.S. AI firm could deter companies from engaging candidly with national-security customers; its characterization of the action as unprecedented is Anthropic’s claim, not an independently established historical finding. Its statement sets out those arguments.
The conflict therefore reaches beyond contract wording. It asks who controls the safety stack, whether a vendor can technically prevent prohibited uses, and whether formal human approval amounts to meaningful control. Legal authority to set procurement conditions would not, on its own, demonstrate that a model can safely perform every lawful task. Conversely, a vendor’s safety policy does not by itself settle what the government may require under a defense contract.
How OpenAI described its alternative
On February 28, OpenAI said it had reached an agreement with the Department of War to deploy advanced AI in classified environments. The company described cloud-only deployment, its control over the safety stack, cleared OpenAI personnel and technical experts remaining in the loop, and limits on autonomous weapons, mass domestic surveillance, and certain high-stakes automated decisions. It also said the systems would not be deployed directly on edge devices capable of powering autonomous weapons. On March 2, OpenAI said the agreement was updated to make the domestic-surveillance restriction more explicit, including restrictions involving commercially acquired personal or identifiable information. These are OpenAI’s descriptions of its own safeguards and contract, not independent verification.
Rank #4
- Runs UniFi Network for full-stack network management
- Manages 30+ UniFi Network devices and 300+ clients
- 1 Gbps routing with IDS/IPS
- Multi-WAN load balancing
- 0.96" LCM status display
OpenAI also said it did not believe Anthropic should be designated a supply-chain risk and urged the government to make comparable terms available to other AI companies. Its agreement did not, by itself, settle Anthropic’s dispute or establish that the two companies’ deployment arrangements were equivalent.
Classified access is not blanket authorization
Anthropic said it was the first frontier AI company to deploy models in U.S. government classified networks and had supported American warfighters since June 2024. Separately, OpenAI announced on February 9, 2026, that ChatGPT was being brought to GenAI.mil, which it described as a secure Department of War enterprise AI platform used by 3 million civilian and military personnel. OpenAI’s announcement describes that platform.
Access to a classified network does not automatically authorize a model to conduct unrestricted surveillance or control weapons. The deployment environment, contract terms, mission authorization, model capabilities, and human oversight all shape what a system may do. A cloud-only arrangement may limit some forms of direct deployment, but it does not answer every question about whether a human’s review is substantive or how safeguards are enforced in practice.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Questions that determine the real-world impact
- Was a final designation issued, and what written determination identifies the legal authority and covered products or services?
- What procurement instruction, if any, tells contractors what they must stop, certify, or replace?
- Does the instruction reach subcontractors, cloud hosts, software vendors, and internal productivity tools, or only specified contract performance?
- Can the government restrict a contractor’s unrelated commercial relationship with Anthropic, and what remedy would apply if it tried?
- Can a human-review requirement ensure meaningful responsibility, rather than merely record a formal approval?
- How would a contractor separate commercial and defense workloads if a model or platform serves both?
Those answers require the operative government documents and, where challenged, legal resolution. Until then, the February 27 announcement should not be treated as proof that every federal agency or defense contractor was legally barred from all Anthropic products.
Why the dispute matters beyond one vendor
The episode tests whether frontier-AI companies can negotiate safety conditions for military work without risking exclusion from the defense supply chain. It also raises a procurement question: if agencies insist on broad “all lawful uses” terms, will vendors accept them, or will the government favor systems with fewer provider-imposed limits? For enterprise and government buyers, the practical lesson is to assess use-case boundaries, deployment mode, human oversight, auditability, data handling, contract restrictions, and the ability to move workloads if policy or vendor access changes. A consumer chatbot subscription is not a substitute for a cleared, government-authorized deployment; contractors should obtain written legal, procurement, cybersecurity, and contract guidance before using an AI service on covered work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
