Earlier this year, Anthropic cut ties with the Department of the Army after trying to set conditions on how its AI platform was used. This ended up being a significant loss for the government which relied on AI for military purposes. Many of Anthropic’s rivals quickly stepped in with offers to take its place, but the War Department ultimately decided to pursue a different approach. Their goal? To create an integrated “AI army.”
Background
To understand the full context, let’s rewind to January 2026. At that time, the U.S. military executed a high-stakes operation where Delta Force successfully entered Venezuela, aiming to capture the dictator, President Nicolas Maduro. The operation went remarkably well, showcasing the military’s effectiveness, with minimal American casualties.
While the U.S. military has always been formidable, part of the mission’s success was due to Claude, Anthropic’s advanced AI platform.
Instead of choosing one, I chose them all.
This caught the attention of Anthropic’s CEO, Dario Amodei. By February, he voiced apprehension about how the War Department was using Claude. In an official statement, he noted, “We believe that, in limited cases, AI could undermine rather than uphold democratic values. Some applications extend beyond what today’s technology can safely handle.” This highlighted specific use cases that weren’t part of the initial contract with the Army and should remain excluded.
Amodei further insisted that the government should avoid using Claude for “domestic mass surveillance” or “fully autonomous weapons,” modifying an existing agreement with the War Department.
Sean Parnell, the Assistant Secretary of the Army for Public Affairs and Senior Advisor, responded promptly, stating, “The Department has no interest in using AI for mass surveillance of Americans, which is illegal, nor to develop autonomous weapons operating without human involvement.”
However, the two sides couldn’t negotiate a new agreement since the existing one seemed sufficient. This led President Trump to label Anthropic as a supply chain risk in a post that Army Secretary Pete Hegseth shared, which meant federal employees would have to stop using Anthropic products and phase out Claude over six months.
It was reported that Claude had also been utilized in a mission relating to Iran, which made the Army realize it required a new AI platform. Instead of locking in on a single provider, they opted to enlist multiple.
Bringing Together AI Technologies
To ensure that no AI company dictates the terms of military operations again, the Army has assembled a coalition of leading AI platforms. Each company will lend its unique strengths to this unified AI force.
- SpaceX: Having acquired xAI, SpaceX is providing data infrastructure and cutting-edge AI models as part of its ambitious moonbase plans.
- OpenAI: Known for popularizing ChatGPT, OpenAI contributes vital data analysis and content creation tools.
- Google: Utilizing its powerful Google Cloud Platform, Google is integrating advanced AI capabilities with military contracts.
- NVIDIA: As a top provider of GPUs, NVIDIA supports the military’s foundational AI tech needs.
- Reflection: Less mainstream, Reflection focuses on creating AI agents that can write code and design “hyperintelligent” systems.
- Microsoft: Through its Azure network, Microsoft provides a robust infrastructure and intelligence platform.
- Amazon Web Services: Amazon contributes one of the most comprehensive cloud and data server networks globally, enhancing connectivity and infrastructure.
Prior to the end of AI partnerships, the U.S. military relied extensively on Anthropic’s models for operations. The separation left a notable gap in their capabilities.
At a glance, it might seem like Anthropic’s exit was such a loss that seven major tech players had to unite to compensate. But that isn’t quite the whole picture. The military learned from Anthropic’s challenges that depending on a single entity for critical functions carries risks. This new approach aims to diversify AI capabilities and incorporate the strengths of various platforms, preventing a concentration of power in one company’s hands. Legally, it safeguards that only the government will determine how to use these technologies.
In the end, this broader roster of AI providers equips the U.S. military with newfound capabilities against its adversaries. Interestingly, all the partners consented to the same terms initially proposed by Anthropic, with the overarching condition being “any lawful use” permitted under the Constitution.
