{"id":24805,"date":"2026-05-01T20:17:08","date_gmt":"2026-05-01T20:17:08","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/24805\/"},"modified":"2026-05-01T20:17:08","modified_gmt":"2026-05-01T20:17:08","slug":"dod-expands-its-classified-ai-work-with-8-companies-excluding-anthropic-amid-ongoing-dispute","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/24805\/","title":{"rendered":"DOD expands its classified AI work with 8 companies \u2014 excluding Anthropic \u2014 amid ongoing dispute"},"content":{"rendered":"

Eight major U.S. technology companies have signed formal agreements to deploy their frontier AI capabilities on the Defense Department\u2019s classified networks \u201cfor lawful operational use,\u201d according to a Pentagon press release<\/a> published Friday.\u00a0<\/p>\n

DOD\u2019s new deals with SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, Amazon Web Services, and Oracle follow a major contract dispute<\/a> between the department and Anthropic that culminated earlier this year over potential ethical constraints that accompany the use of AI in warfare and for national surveillance.<\/p>\n

\u201cIntegrating secure frontier AI capabilities into the Department\u2019s Impact Level 6 (IL6) and Impact Level 7 (IL7) network environments will streamline data synthesis, elevate situational understanding, and augment warfighter decision-making in complex operational environments,\u201d officials wrote in Friday\u2019s press release.<\/p>\n

DOD relies on its \u201cImpact Level\u201d<\/a> classification system to categorize data and securely authorize cloud-based hosting environments and services.\u00a0<\/p>\n

Different federal agencies abide by different security protocols. For DOD, IL6 marks a rigid compliance standard that is required to process classified data for cloud-based defense workloads. Beyond that is IL7 \u2014 the most stringent security classification, which covers cloud computing environments designed to handle top secret, highly sensitive, or critical national security information.<\/p>\n

The Pentagon\u2019s press release notes that the eight companies \u201cwill provide resources to deploy their capabilities on both IL6 and IL7 environments.\u201d<\/p>\n

However, it\u2019s unclear whether all those tech giants have already been fully authorized to deploy their capabilities on DOD\u2019s classified networks at this time, or if some are on expedited paths to get there.\u00a0<\/p>\n

Pentagon spokespersons did not respond to DefenseScoop\u2019s request for more information on Friday.<\/p>\n

This effort supports the Pentagon\u2019s AI acceleration strategy \u201cby enabling new capabilities across its three core tenets of warfighting, intelligence, and enterprise operations,\u201d officials wrote in the press release, and the underpinning agreements are expected to \u201caccelerate the transformation toward establishing the United States military as an AI-first fighting force.\u201d<\/p>\n

Frontier AI refers to the most advanced, large-scale foundational models that are now pushing the boundaries of machine intelligence. While these powerful capabilities hold massive potential to transform military operations, the still-emerging models also pose serious risks<\/a> to humanity that can range from ethical dilemmas to existential threats.<\/p>\n

Last summer, Pentagon leaders unveiled individual contracts<\/a> with OpenAI, Anthropic, Google and xAI \u2014 each worth up to $200 million \u2014 for \u201cfrontier AI projects.\u201d Then, in December, DOD launched the enterprise-wide generative AI platform<\/a>, GenAI.mil.<\/p>\n

That hub is designed to<\/a> deploy secure and reliable genAI tools<\/a> to all service members and civilian users for tasks and data at IL5, which encompasses the highest authorization level granted to environments built to store and process controlled unclassified information (CUI).<\/p>\n

\u201cOver 1.3 million Department personnel have used the platform, generating tens of millions of prompts and deploying hundreds of thousands of agents in only five months,\u201d Friday\u2019s press release stated.<\/p>\n

Of DOD\u2019s original GenAI.mil partners, Anthropic\u2019s models were the only ones at the time to also be integrated into the department\u2019s classified workflows, via a partnership with Palantir.\u00a0<\/p>\n

But tensions between Anthropic and the Pentagon heightened<\/a> in early 2026, reportedly stemming from disagreements over whether and how the military was applying Claude models in certain operations. DOD leaders subsequently moved to blacklist Anthropic as a \u201csupply chain risk\u201d \u2014 a designation that is typically reserved for foreign adversaries.\u00a0<\/p>\n

Anthropic sued the department in federal court and this high-stakes scuffle remains in active litigation.<\/p>\n

Since the initiation of the GenAI.mil effort, the Pentagon has come to recognize that \u201cit\u2019s irresponsible to be reliant on any one partner,\u201d its CTO and Undersecretary of Defense for Research and Engineering Emil Michael told CNBC<\/a> on Friday.\u00a0<\/p>\n

\u201cAnd we learned that that one partner didn\u2019t really want to work with us in the way we wanted to work with them. We went out and made sure that we had multiple different providers \u2014 both open source, which is [a new] effort here at the department, and the proprietary model companies, and the infrastructure companies like Microsoft and AWS \u2014 and got them to agree to sign up to work with us on classified networks to make sure we had diversity supply,\u201d Michael said. \u201cWe had the best of the best, and we had multiple different paths with open source and proprietary.\u201d<\/p>\n

In response to questions from DefenseScoop on Friday, Center for Security and Emerging Technology (CSET) Senior Research Analyst Lauren Kahn said DOD\u2019s move to expand its industry-enabled classified AI work marks \u201ca step in the right direction \u2014 it is necessary and, frankly, inevitable.\u201d<\/p>\n

She also called it \u201clong overdue.\u201d<\/p>\n

\u201cUltimately, the Pentagon is a world of paper \u2014 the world\u2019s largest bureaucracy \u2014 and this could dramatically facilitate work, much of which involves printing read-aheads, filling out rigid forms, and formatting talking points just right. Anthropic was the first company on the high-side, which paved the way for these other companies,\u201d she explained. \u201cThese models all have different strengths and are evolving and updating at lightspeed, so it\u2019s important to have access to a variety and let Pentagon users shop around \u2014 giving them access to the same capabilities civilians have in their everyday lives, and the same way they do now on the unclassified side.\u201d<\/p>\n

Prior to CSET, Kahn was a policy advisor for force development and emerging capabilities in the Office of the Undersecretary of Defense for Policy at DOD.<\/p>\n

In her view, the announcement \u201csmartly highlights upfront\u201d that these systems will be used for \u201clawful operational use,\u201d thus acknowledging some of the anxieties circulating about frontier AI in the wake of the recent events.\u00a0<\/p>\n

It\u2019s not covered in the press release, but Kahn said the next step the Pentagon must pursue \u201cto have any real effect at scale\u201d will involve ensuring operators and DOD users are trained against things like automation bias, the tendency to overdelegate to machines \u2014 and that they begin to understand the benefits, pitfalls, and limitations of these systems.\u00a0<\/p>\n

\u201cHaving access to multiple models, other than hedging against vendor lock-in, actually helps accelerate that learning, because users can directly compare responses, accuracy and speed, and start to appreciate that not all these systems work the same way,\u201d Kahn told DefenseScoop.<\/p>\n

\t\t\t\t\t\"Brandi<\/p>\n

\n\t\t\tWritten by Brandi Vincent
\n\t\t\tBrandi Vincent is a Senior Reporter at DefenseScoop, where she reports on disruptive technologies and associated policies impacting Pentagon and military personnel. Prior to joining SNG, she produced a documentary and worked as a journalist at Nextgov, Snapchat and NBC Network. Brandi grew up in Louisiana and received a master\u2019s degree in journalism from the University of Maryland. She was named Best New Journalist at the 2024 Defence Media Awards.\t\t<\/p>\n","protected":false},"excerpt":{"rendered":"Eight major U.S. technology companies have signed formal agreements to deploy their frontier AI capabilities on the Defense…\n","protected":false},"author":2,"featured_media":24806,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[24,1395,53,25,111,322,11082,16777,11085,11030,223,132,16778,320,58,157,2163,16779,658],"class_list":{"0":"post-24805","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-ai","8":"tag-ai","9":"tag-amazon-web-services","10":"tag-anthropic","11":"tag-artificial-intelligence","12":"tag-artificial-intelligence-ai","13":"tag-aws","14":"tag-cdao","15":"tag-classified-systems","16":"tag-emil-michael","17":"tag-genai-mil","18":"tag-generative-ai","19":"tag-google","20":"tag-lauren-kahn","21":"tag-microsoft","22":"tag-nvidia","23":"tag-openai","24":"tag-oracle","25":"tag-reflection","26":"tag-spacex"},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/24805","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=24805"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/24805\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/24806"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=24805"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=24805"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=24805"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}