The United States administration has added four new artificial intelligence companies to its roster of preferred suppliers for military and classified operations. The Pentagon signed agreements with Microsoft, Reflection AI, Amazon, and Nvidia, allowing their products to be deployed in classified contexts. These companies join OpenAI, xAI, and Google as vendors the Department of Defense can use for what it terms any lawful use.
Dispute Over Terms and Contract Cancellation
The phrase any lawful use became a central point of contention between the administration and Anthropic AI. CEO Dario Amodei argued that the language would permit the government to use Anthropic technologies for domestic surveillance and autonomous weapons development, areas the company wanted to exclude. The Pentagon subsequently canceled a $200 million contract with Anthropic. The company filed a lawsuit, claiming substantial revenue losses from both the government and other clients influenced by the decision.
The Trump administration designated Anthropic as a supply chain risk, marking the first time a U.S. based company received such a classification. Government statements later characterized Anthropic as a woke company.
Rationale for Expanded Supplier Base
The Pentagon stated that the new agreements aim to build an architecture that prevents AI vendor lock in and ensures long term flexibility for the Joint Force. The technologies will give warfighters the tools they need to act with confidence and safeguard the nation against any threat. The AIs will be used for Impact Level six, covering secret data, and Impact Level seven, the most highly classified materials. The goal is to create what the Pentagon describes as an AI first fighting force.
Currently, the Pentagon uses generative AI primarily for non classified tasks such as document drafting, summarization, and research within various defense departments. The new suppliers will also help streamline data synthesis, elevate situational understanding, and augment warfighter decision making in complex operational environments. It remains unclear whether these deployments include domestic operations inside U.S. borders.
Implications for Military and National Security
Expanding the supplier base makes the military and security forces more resilient to individual vendor decisions. By broadening their technological foundation, the personal preferences of company leaders become less consequential. Google and Amazon previously faced internal protests when employees objected to their companies technology being used in weaponry and warfare applications.
Anthropic Claude AI was previously used on classified material through Palantir Maven toolset, a role new signees may fill. However, reports indicate that Anthropic Mythos model is currently used by the National Security Agency for cyber warfare and defense capabilities. Approximately 40 organizations worldwide are assessing Mythos, with only 12 named publicly. Among the unnamed 28 are believed to be the United Kingdom MI5 and the U.S. National Security Agency.
Possible Reversal of Stance
According to Axios, the administration may be reconsidering its recent public position toward Anthropic. A White House source told the outlet that officials are trying to find ways to save face and bring them back in. Anthropic Claude coding model is reportedly still in use by U.S. government security organizations throughout the ongoing dispute.
The White House stated that the U.S. government continues to proactively engage across government and industry to protect our country and the American people, including by working with frontier AI labs.
Looking Forward
No official timeline has been provided for potential reconciliation with Anthropic. The legal proceedings initiated by the company remain pending. The government expanded supplier network is expected to become operational in phases, with integration into classified environments proceeding over the coming months. Further developments will likely depend on court outcomes and administrative policy adjustments.
