五角大楼与OpenAI、谷歌和英伟达达成了机密的人工智能交易—但不是人类| The Verge

Pentagon strikes classified AI deals with OpenAI, Google, and Nvidia — but not Anthropic

2026年5月1日 22:09 The Verge AI 国际资讯关注 3 分钟阅读 414 字归档：2026年5月2日 02:14 查看原文 →

发布算力

Pentagon strikes classified AI deals with OpenAI, Google, and Nvidia — but not Anthropic | The Verge

导读

五角大楼已与七家人工智能公司达成协议，允许OpenAI、谷歌、微软、英伟达和其他公司处理机密信息。

The Pentagon has struck deals with OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk's xAI, and the startup Reflection, allowing the agency to use their AI tools in classified settings, according to an announcement on Friday. At the same time, the Defense Department has left out Anthropic - which it previously used

原文快照

站内保留一份可阅读的正文副本；如抓取失败，则保留摘要和原文链接。

Pentagon strikes classified AI deals with OpenAI, Google, and Nvidia — but not Anthropic

The flurry of deals leaves out Anthropic, which the department previously used for handling classified information.

The Pentagon has struck deals with OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk’s xAI, and the startup Reflection, allowing the agency to use their AI tools in classified settings, according to an announcement on Friday. At the same time, the Defense Department has left out Anthropic — which it previously used for classified information — after declaring it a supply-chain risk.

This builds upon deals with OpenAI and xAI, which have already reached agreements with the Pentagon for the “lawful” use of their AI systems. A report from The Information suggests Google has struck a similar agreement. As noted by The Wall Street Journal, Microsoft and Amazon already have “deep relationships with the Pentagon,” while contracts with Nvidia and Reflection are new.

And while Anthropic had a $200 million deal to handle classified materials for the Pentagon, it refused to loosen “red lines” around mass domestic surveillance and fully autonomous weapons for the agency, leading to a dispute that banned the AI startup’s products from the federal government. Anthropic sued the federal government in response and won a temporary injunction.

Anthropic doesn’t want its AI killing people unsupervised. The Pentagon isn’t happy.

Google and Pentagon reportedly agree on deal for ‘any lawful’ use of AI

Emil Michael, the Defense Department’s chief technology officer, told CNBC on Friday that Anthropic is still a supply chain risk, but called its powerful security model, Mythos, a “separate national security moment,” adding that “we have to make sure that our networks are hardened up, because that model has capabilities that are particular to finding cyber vulnerabilities and patching them.”

In its announcement, the Pentagon says the agreements with the seven AI companies will allow for the “lawful operational use” of their systems, “establishing the United States military as an AI-first fighting force.”

Emma Roth

More in: Anthropic vs. the Pentagon: killer robots, mass surveillance, and red lines