Anthropic secured a federal court ruling vacating the Pentagon's supply-chain blacklist1 and released a framework that lets AI agents operate physical lab and factory equipment5,13. ANALYSIS Taken together, the moves trace a company pushing outward on multiple strategic axes at once: legal and technical.
The court ruling does more than restore Anthropic's access to government contracts. U.S. District Judge Rita Lin found that the Defense Department's designation of Anthropic as a supply-chain risk constituted "unlawful retaliation in violation of the First Amendment". "The empty invocation of national security is not a blank check to punish and retaliate against government critics," Lin wrote in a 59-page decision14,15. Lin also ruled the designation was "arbitrary and capricious" and that Anthropic was denied due process under the Fifth Amendment.
Anthropic argued the designation could cost the company "billions of dollars in lost business and reputational harm". Lin noted the government's own behavior undercut its stated rationale, observing that the Pentagon continued pursuing work with Anthropic even after the blacklist: "None of that is consistent with a genuine fear that Anthropic is a saboteur who would poison its software to harm national security".
The legal fight originated in the Pentagon's effort to bring Claude onto an internal platform called GenAI.mil4. Anthropic insisted that two areas remain off-limits: fully autonomous weapons and mass surveillance of American citizens2. The Defense Department sought to use Claude for "all lawful purposes," including the most sensitive military and intelligence applications. When Anthropic held its red lines, Defense Secretary Pete Hegseth labeled the company a supply-chain risk and ordered all federal agencies to stop working with it. Lin found the government's actions were "based on a desire to make a public example out of Anthropic for its 'arrogance' in criticizing the government"11.
Lin also flagged an internal contradiction in the government's approach, pointing to Hegseth's proposition to apply the Defense Production Act to Anthropic, "which would" compel cooperation rather than sever it. ANALYSIS That detail underscores a tension in the administration's posture: one arm of the government moved to cut off Anthropic while another sought to compel its participation, an inconsistency the court treated as evidence against a genuine national-security rationale.
The government is expected to appeal the ruling. Anthropic is also fighting a separate Pentagon designation under a different statute in the D.C. Circuit. "We remain focused on working productively with the government to harness AI for our national security so all Americans benefit from this technology," an Anthropic spokesperson said.
On the product side, Anthropic released a research preview of the Model Hardware Standard, a framework that enables AI agents to operate physical devices such as microscopes, robotic arms, and lasers. MHS is a set of standardized drivers designed to let AI agents interface with and control arbitrary devices3. The standard works on any device with a programmable interface and is independent of the base model type. Anthropic said MHS can shorten hardware integration work from "weeks or months" to "hours or minutes". MHS enables AI agents to operate lab and manufacturing instruments in tandem and perform complex tasks ranging from routine drug discovery experiments to laser calibration on a quantum computer.
The development began in collaboration with HHMI. Anthropic Technical Staffer Alek Kemeny said the MHS effort was inspired by observing neuroscientist Arco Bast work through an experiment on memory formation in the brain at the HHMI Janelia Research Campus in Ashburn, Virginia. Kemeny recalled thinking, "This idea could be used to have AI run any science experiment in the world".
MHS provides a common interface and common format for data sharing between devices, letting them communicate across a network "without needing a bespoke 'translator' program in between". The framework aims to help researchers and engineers execute autonomous, round-the-clock workflows with minimal human intervention. MHS does not yet work with hardware that lacks a programmable interface. Anthropic has so far made the technology accessible only to a limited number of users. Anthropic is sharing an early version of MHS with partners to build out safety evaluations before making it open source.
ANALYSIS The MHS research preview marks a concrete step toward extending agentic AI beyond software-only environments into physical laboratory and manufacturing workflows. The requirement for a programmable interface on target devices defines the current boundary of the standard's applicability.
Separately, Anthropic is in talks with chip startup MatX to speed up chip design8. Details of those discussions remain limited to the headline-level report.
"We welcome the court's ruling that this supply chain risk designation was unlawful," the Anthropic spokesperson told CNBC.