The Pentagon wanted partners who could move fast, not companies that would slow them down
In a quiet but consequential shift, the Pentagon has formalized classified AI partnerships with seven of the technology industry's most powerful firms, including Nvidia, Microsoft, and Amazon Web Services, drawing the machinery of commercial innovation deeper into the architecture of national defense. The agreements reflect a long-building conviction within the Defense Department that military advantage in the coming era will be shaped not on traditional battlefields alone, but in the speed and sophistication of artificial intelligence. Notably absent from these arrangements is Anthropic, a company that has staked its identity on the careful, cautious development of AI — and whose exclusion invites reflection on what values, and what trade-offs, now govern the relationship between technology and power.
- The Pentagon has signed classified AI contracts with seven major tech firms, marking one of the most significant expansions of military-commercial AI collaboration in recent memory.
- Anthropic's conspicuous exclusion from the deals has unsettled observers, raising urgent questions about whether a commitment to AI safety is now a liability in defense procurement.
- The selected companies — anchored by Nvidia's hardware dominance, Microsoft's enterprise reach, and AWS's cloud infrastructure — represent the current power centers of the AI industry, now formally enlisted in classified military work.
- The classified structure of the contracts means the public will have little visibility into what systems are being built, for what purposes, and with what safeguards — a transparency gap that critics are already noting.
- The deals are sending a signal across the tech sector: defense work is becoming a defining business frontier, and how companies position themselves on safety and military engagement may shape their futures in profound ways.
The Pentagon has moved quietly but decisively to deepen its relationship with the technology industry, signing classified agreements with seven major companies to develop and deploy artificial intelligence for military purposes. Nvidia, Microsoft, and Amazon Web Services are among the named partners, with the remaining four undisclosed. Together, they represent the current center of gravity in commercial AI — from chip manufacturing to cloud infrastructure to enterprise software.
What gives this moment its particular weight is not only the scale of the commitment, but the notable absence of Anthropic. The San Francisco-based company has built its reputation around responsible AI development and has been among the most prominent voices urging caution before deploying advanced systems in high-stakes environments. Its exclusion has prompted immediate questions about the Pentagon's selection criteria — whether existing defense relationships, established security clearances, or simply a preference for speed over caution shaped who was invited in.
The contracts are structured around classified networks, meaning the work will involve sensitive military information and systems shielded from public scrutiny. The Defense Department has framed accelerating AI adoption as a strategic imperative, driven in large part by concern about falling behind China in AI capabilities. These agreements signal a belief that private industry's innovation velocity is essential to closing that gap.
The broader implications extend well beyond the Pentagon's immediate needs. By making defense work a central business opportunity, these partnerships may quietly reshape how AI companies set their development priorities — and how seriously they weigh safety considerations against the pull of lucrative government contracts. Whether others follow Anthropic's more selective posture, or whether this moment sets a new norm for tech-military collaboration, remains the defining question now unfolding.
The Pentagon has quietly moved to deepen its ties with the technology industry's largest players, signing classified agreements with seven major companies to develop and deploy artificial intelligence systems for military use. The deals, which involve Nvidia, Microsoft, and Amazon Web Services among others, represent a significant expansion of how the Defense Department plans to integrate AI into its operations—from logistics and intelligence analysis to weapons systems and strategic planning.
What makes this round of contracting notable is not just its scope but whom it excludes. Anthropic, one of the most prominent AI safety-focused companies in the sector, was notably left out of the agreements. The San Francisco-based firm, which has positioned itself as a leader in responsible AI development and has been vocal about the risks of deploying advanced systems without adequate safeguards, did not receive a contract. The omission raises immediate questions about how the Pentagon evaluated its potential partners and what criteria determined who would gain access to classified military work.
The seven companies that secured deals represent the current center of gravity in the AI industry. Nvidia dominates the hardware side—its processors power most large language models and AI training operations worldwide. Microsoft has embedded AI throughout its enterprise software and cloud services, and has invested heavily in OpenAI. Amazon's cloud division, AWS, provides infrastructure for countless AI applications. The other four companies in the agreement have not been publicly named in available reporting, though the Pentagon's announcement suggests they represent additional significant players in the commercial AI space.
These contracts are structured around classified networks, meaning the work will involve handling sensitive military information and developing systems that operate on secure, compartmentalized systems. The Pentagon has been increasingly concerned about falling behind in AI capabilities relative to potential adversaries, particularly China, and has made accelerating AI adoption a strategic priority. These agreements signal that the department believes it needs the innovation velocity and technical expertise of private industry to achieve that goal.
The exclusion of Anthropic is particularly striking given the company's prominence in recent AI policy debates. Anthropic has been among the most vocal advocates for AI safety standards and has publicly cautioned against deploying advanced systems without robust testing and oversight. Some observers have suggested that the company's emphasis on safety and its willingness to decline lucrative contracts that conflict with its stated values may have made it less attractive to Pentagon planners focused on rapid capability development. Others have speculated that the selection process may have favored companies with existing defense relationships and established security clearances.
The Pentagon has not publicly detailed the specific capabilities these companies will develop or the total value of the contracts. The classified nature of the work means much of what these partnerships will produce will remain hidden from public view. What is clear is that the agreements represent a significant vote of confidence in commercial AI technology and a commitment to integrating it deeply into military operations in ways that were unthinkable just a few years ago.
These partnerships will likely accelerate the pace at which AI systems move from research labs into operational military use. They also signal to the broader tech industry that defense work is becoming a central business opportunity, potentially influencing how companies structure their AI development priorities and which safety considerations they prioritize. The coming months will reveal whether other companies follow Anthropic's apparent path of selective engagement with defense work, or whether the Pentagon's deals set a new standard for tech-military collaboration.
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter that Anthropic wasn't included? They're one company among many.
Because Anthropic has been the loudest voice saying we should slow down and think carefully about what we're building. If the Pentagon wanted that perspective in the room, they would have asked for it. The fact that they didn't suggests something about what they're actually prioritizing.
You mean speed over safety?
Not necessarily. But it suggests they're optimizing for companies that can move fast and integrate with existing defense infrastructure. Anthropic has made a point of being selective about its partnerships. That selectivity may have cost them.
Is this a problem?
It depends on what you think the Pentagon should be doing with AI. If you believe military AI systems need the most rigorous safety testing possible, then yes—excluding the company most vocal about that is concerning. If you believe the U.S. needs to move quickly to match adversary capabilities, then it makes sense to work with partners who won't slow you down with extra caution.
What happens now?
These seven companies will build classified systems we won't see for years, maybe decades. Other tech companies will watch and decide whether to pursue defense contracts. And Anthropic will have to decide whether to eventually engage with military work or remain on the sidelines.
Could Anthropic push back?
They could speak publicly about their concerns, but that's a risky move when you're already excluded from a major opportunity. The real leverage belongs to the Pentagon right now.