US War Dept Taps 7 AI Powerhouses-Secret Models Inside Classified Networks!

by

US Department of War Signs AI Agreements With 7 Top Tech Companies

On Friday, the U.S. Department of War partnered with seven major American tech and infrastructure companies. They signed agreements to use the most advanced AI models on secure, classified networks.

These agreements include SpaceX, OpenAI, Google, NVIDIA, Reflection AI, Microsoft, and Amazon Web Services. They allow these companies’ artificial intelligence systems to function within secure government environments – specifically Impact Levels 6 and 7 – for any legal purpose.

Inside the Department of War’s AI Agreements

On May 1st, the Department’s Chief Technology Officer unveiled a new initiative, describing it as part of a larger effort to make artificial intelligence a central focus within the War Department. These new AI models, labeled IL6 and IL7, are designed for handling highly classified information and will work with sensitive intelligence and operational data.

The Office of the Under Secretary of War for Research and Engineering announced this new project as part of its larger effort to prioritize artificial intelligence, stating it’s a step toward making AI central to the department’s work.

According to officials, the Department deliberately chose to work with many different US companies. This approach is meant to prevent reliance on a single vendor and allow flexibility in using both proprietary and open-source technologies.

NVIDIA will contribute its Nemotron family of open-source models. Additionally, Reflection AI, a startup supported by NVIDIA and created by ex-Google DeepMind scientists, will provide other openly available models.

Google is making its Gemini AI models available for use by governments for legitimate purposes, and SpaceX is anticipated to provide the necessary infrastructure for xAI’s Grok AI models.

Microsoft and AWS keep their roles as cloud and infrastructure backbone for the rollout.

The Department of Defense announced today that it has partnered with seven major artificial intelligence companies – SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, and Amazon – to integrate their most advanced AI technologies onto the Department’s secure networks.

— Department of War CTO (@DoWCTO) May 1, 2026

The Department’s new GenAI.mil platform is proving popular internally, with over 1.3 million users and tens of millions of requests submitted in just five months, as of May 1st.

Anthropic Sits Out After Guardrail Standoff

Anthropic is not on the list. In February, Defense Secretary Pete Hegseth identified the company as a potential supply-chain problem because they wouldn’t lift limitations on the development of self-governing weapons and widespread internal monitoring.

Defense Department spokesperson Sean Parnell stated that the department will make its own operational decisions and won’t allow any company to control those choices.

A federal judge later blocked enforcement of the ban, and the legal fight continues.

OpenAI chose a more focused approach compared to its competitors. The company stated that its agreement with the Department of Defense upholds three key principles:

  • Its models cannot be used for mass domestic surveillance,
  • Cannot direct autonomous weapons, and
  • Will keep their safety guardrails in place.

We’ve reached an agreement with the Department of War to implement our AI models on their secure network. Throughout our discussions, they consistently prioritized safety and demonstrated a strong commitment to working with us for the best results. This marks a significant step towards broader AI deployment and, crucially, maintaining safety standards.

— Sam Altman (@sama) February 28, 2026

Other firms accepted broader “any lawful purpose” language without those public carve-outs.

Open-Source Push Sets the Tone for What Comes Next

These investments support the Department’s plan, released earlier in 2026, to use flexible, open-source systems for military operations, intelligence gathering, and general business functions.

As a crypto investor, I’m really encouraged to hear they’re prioritizing solutions built by local companies, focusing on open-source and adaptable technologies, and emphasizing quick development and testing. It sounds like they’re avoiding being locked into proprietary systems, which is a smart move for long-term innovation and control.

We’ll be closely watching which AI models successfully implement IL6 first, and whether OpenAI’s safety measures remain effective as more sensitive tasks are used with them.

Read More

2026-05-01 17:51