OpenAI has signed a major contract with the Pentagon and is now reportedly seeking a similar deal with NATO. The move marks a sharp turn for a company that once promised to keep its artificial intelligence away from military use. The decision has triggered resignations, user protests, and a fierce debate about the future of AI in defense.
The Pentagon deal was announced on February 28, 2026. It is worth 200 million dollars and focuses on developing advanced AI tools for national security. OpenAI will deploy its technology within classified military networks. The company says the work covers warfighting and enterprise needs, with completion expected by July 2026.
Source: Reuters
From AI for Humanity to AI for Defense
The shift is dramatic. For years, OpenAI positioned itself as a company building AI for the benefit of all humanity. Its charter explicitly stated that the technology should be used for peaceful purposes. Now the company is actively pursuing military contracts and defending its decision to work with the Pentagon.
CEO Sam Altman acknowledged the change in an internal meeting. He admitted the deal was rushed and that the optics were not good. But he insisted the move was necessary to protect both the company and the broader AI industry. According to the Wall Street Journal, Altman told staff that the backlash was painful but that a good relationship between the government and AI companies is critical.
Source: Wall Street Journal
Altman also said he disagreed with the government’s decision to label rival Anthropic as a supply chain risk. He called it a very bad decision and said he hoped the government would reverse it. This suggests OpenAI sees itself as mediating between the industry and the military rather than simply serving as a contractor.
sextbots
Source: Wall Street Journal
Employee Revolt and High-Profile Resignations
The Pentagon deal caused immediate internal turmoil. Caitlin Kalinowski, OpenAI’s head of hardware and robotics, resigned on March 7, 2026. She had joined the company in November 2024 from Meta, where she led augmented reality hardware development. In a ai nudifier public post, she wrote that the decision was not easy but that she could not stay after the rushed announcement without clear guardrails.
Her resignation sent a strong signal. Kalinowski was not a junior employee. She was a senior leader with deep experience in hardware safety and privacy. Her departure showed that even top executives were uncomfortable with the direction the company was taking.
Other employees also spoke out. Leo Gao, who works on AI alignment and safety, criticized the company on social media. He said the contract allowed the Pentagon to use OpenAI technology for all lawful purposes and accused the company of adding window dressing to make it seem like there were real restrictions.
The backlash was not limited to employees. Users responded quickly. ChatGPT uninstalls surged by 295 percent after the deal was announced. Some critics started a campaign to persuade users to switch to Anthropic’s Claude chatbot. The campaign had an effect. Claude briefly surpassed ChatGPT as the most downloaded free app in Apple’s App Store.
The NATO Connection
While OpenAI was dealing with internal and external backlash, reports emerged that the company was already looking for its next military customer. According to Reuters, OpenAI is considering a contract to deploy its AI technology on NATO’s unclassified networks.
The discussions were first reported by the Wall Street Journal, which said OpenAI had explored providing services across NATO’s classified systems. A company spokeswoman later clarified that CEO Sam Altman had misspoken in an internal meeting. She said the potential contract relates only to NATO’s unclassified networks, not classified ones.
NATO has been actively pursuing AI integration. In July 2024, the alliance released a revised artificial intelligence strategy. The strategy emphasizes responsible use, interoperability, and collaboration with industry. It also stresses the importance of integrating AI into defense and security applications while maintaining human oversight.
Source: NATO AI Strategy
For OpenAI, a NATO deal would expand its military footprint beyond the United States. The alliance includes 32 member countries. A contract with NATO could mean OpenAI technology is used across multiple allied defense networks. This would represent a major commercial opportunity but also a significant expansion of the company’s involvement in global security.
OpenAI vs Anthropic: Two Different Paths
The contrast between OpenAI and Anthropic could not be clearer. While OpenAI rushed to sign the Pentagon deal, Anthropic refused a similar contract. The Trump administration responded by labeling Anthropic a supply chain risk. Anthropic CEO Dario Amodei has publicly opposed using AI for mass domestic surveillance or fully autonomous weapons.
This divergence has created a split in the AI industry. OpenAI is now seen as the company willing to work with the military under almost any terms. Anthropic is seen as the company that prioritizes ethical boundaries over commercial opportunity. The market is responding. Users are switching to Claude. Investors are watching closely.
OpenAI has tried to defend its position by publishing details of its contract. The company says the agreement includes three main safety red lines. First, OpenAI technology cannot be used for mass domestic surveillance of US citizens. Second, it cannot be used to operate autonomous weapons systems. Third, it cannot be used for high-risk automated decisions like social credit systems.
The company also says the deployment is cloud-based, running on OpenAI’s controlled security stack. OpenAI staff will be involved in monitoring use. The contract references existing laws that limit surveillance and autonomous weapons. If the government violates the terms, OpenAI says it can terminate the contract.
However, critics argue these safeguards are not enough. They point out that the Pentagon can change its own policies at any time. The term mass surveillance is not clearly defined. The contract does not prevent military intelligence agencies from using commercially available data for analysis. And the cloud-based deployment does not eliminate the risk of misuse.
The Broader Implications
This dispute is about more than one contract. It raises fundamental questions about who should control AI and how it should be used.
Some argue that AI companies should work with democratic governments to ensure the technology is used responsibly. They say that if American companies refuse to work with the Pentagon, authoritarian regimes will fill the gap. The US military needs the best tools to maintain national security. AI is just another technology that can serve peaceful or military purposes depending on how it is used.
Others argue that AI is different. They say that large language models and other advanced AI systems are general-purpose technologies that can be used for surveillance, propaganda, and automated warfare in ways that previous technologies could not. Once a company opens the door to military use, there is no going back. The commercial pressure to expand military contracts will only grow.
Altman himself has acknowledged the tension. He said he believes in the democratic process and that elected leaders should have the power to decide how AI is used. But he also said he is terrified of a world where AI companies act like they have more power than the government. He added that he would also be terrified of a world where the government decided mass domestic surveillance was okay.
What Happens Next
The situation is still unfolding. OpenAI has amended its Pentagon deal to add clearer language about surveillance restrictions. The company is pushing for the same contract terms to apply to all AI companies working with the military. It says this would level the playing field and prevent any single company from being singled out.
Anthropic is challenging the supply chain risk designation in court. The company has filed a lawsuit against the Pentagon, arguing that the label is unfair and politically motivated. The outcome of this case could affect how the government treats AI companies that refuse military contracts.
For users and developers, the choice is becoming clearer. OpenAI offers cutting-edge models and deep government integration. Anthropic offers strong ethical commitments and privacy protections. The market is fragmenting along these lines. Some enterprises are reviewing their OpenAI contracts. Others are switching to Claude or other alternatives.
Final Thoughts
OpenAI’s pursuit of military contracts represents a watershed moment for the AI industry. The company that once promised to build safe AI for humanity is now building AI for the Pentagon and seeking deals with NATO. The transformation is complete.
The backlash from employees and users shows that this shift is not universally accepted. The resignations, the uninstalls, and the public criticism all signal that many people in the AI community are uncomfortable with the militarization of the technology they helped create.
The deeper question is whether any AI company can resist the pull of military contracts. The money is too big. The pressure from governments is too strong. The technology is too powerful. In a world where AI is the new oil, the military is the biggest customer. OpenAI may be the first major AI company to openly embrace this reality. It will not be the last.
