Select Language:
The Pentagon announced on Friday that it has formed agreements with seven AI companies to integrate their advanced systems into the Department of Defense’s classified networks. This move aims to expand the variety of AI providers supporting military operations. Notably absent from this list is Anthropic, which has been at odds with the Pentagon over strict guidelines for its AI tools’ use in military applications.
Earlier this year, the Pentagon classified Anthropic as a supply chain risk and prohibited its AI from being used by the Department and its contractors. The companies set to be integrated include SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, and Amazon Web Services—many of which already collaborate with the Pentagon. These will be incorporated into the Impact Levels 6 and 7 network environments, providing broader access to their AI products across more military units.
This expansion allows personnel to leverage AI for planning, logistics, targeting, and numerous other functions—helping streamline complex operations and accelerate decision-making processes. The Pentagon emphasized that this strategy helps avoid overreliance on any single vendor, particularly in relation to Anthropic.
Despite orders to phase out Anthropic’s tools within six months, some Pentagon staff, former officials, and IT contractors familiar with military AI expressed reluctance to relinquish its superior tools. They view Anthropic’s technology as more effective than alternatives currently available.
AI has increasingly become a critical resource for the U.S. military. The Pentagon highlighted that its main AI platform has been used by over 1.3 million Department of Defense personnel within just five months of launching.
In addition, Google, which already serves the Pentagon, recently signed an agreement allowing the department to utilize its AI models for classified operations, according to a source familiar with the matter.
Department CTO Emil Michael told CNBC that Anthropic remains a supply chain risk, but he also noted that Mythos—Anthropic’s AI model with advanced cyber capabilities—constitutes a separate national security concern, especially given its potential to enhance hacking activities. While many companies and private entities have access to Mythos for cybersecurity purposes, it’s unclear if the Pentagon participates in that program.
Last week, President Donald Trump suggested that the Biden administration’s view of Anthropic might be softening, raising the possibility that the company could be removed from the Pentagon’s blacklist. However, the ongoing tensions underscore the importance of diversifying AI sources for the military, opening new opportunities for smaller defense tech startups.
ChatGPT
Perplexity
Gemini AI
Grok AI
