The real war AI—Claude is just the underlying layer, Palantir is helping the US military fight wars.

```

The ongoing dispute between the Pentagon and Anthropic is drawing public attention to a previously seldom-examined area: how AI technology is actually operationalized in U.S. military operations.

According to an article in Wired magazine last week, Palantir is embedding AI chatbots into the core layer of U.S. military combat systems, and Anthropic's Claude is merely one of several interchangeable underlying models.

As revealed by Wired's review of Palantir software demonstrations, public documents, and Pentagon records, Palantir has built an AI-assisted system covering intelligence analysis, target identification, strike plan generation, and operational route planning, deploying it across multiple levels of military command. Claude’s role is as the "language engine" for the chatbot, not as the system itself.

The exposure of this architecture comes amid intensifying legal conflicts between Anthropic and the Trump administration. Anthropic filed two lawsuits this week, accusing the Pentagon's decision to list its products as a "supply chain risk" as unlawful retaliation. Meanwhile, Claude has been continuously used in U.S. military overseas operations—including the Iran conflict—and reportedly played a key role in the military action leading to the capture of Venezuelan President Nicolás Maduro.

Palantir's Military System Landscape

Palantir's cooperation with the Pentagon dates back to 2017. Since then, Palantir has served as the main contractor for "Project Maven" (the Algorithmic Warfare Cross-Functional Team)—the Department of Defense's core project for deploying AI in war settings.

Palantir’s central product for the project is called Maven Smart System (Maven for short), managed by the National Geospatial Intelligence Agency. The U.S. Army, Air Force, Space Force, Navy, Marine Corps, and U.S. Central Command responsible for Iran military operations all have access to the system. Pentagon Chief Digital and AI Officer Cameron Stanley recently said at a Palantir conference that Maven is being deployed "department-wide."

According to publicly released military assessment documents, Maven can apply "computer vision algorithms" to images taken by space-based assets like satellites, automatically identifying targets that may belong to "enemy systems." The system's built-in visualization tools can tag potential strike targets and "nominate" them for ground or air strikes. A feature called the "AI Asset Mission Recommender" can suggest which bombers and munitions to assign to which targets. Maven also handles the transmission of "target intelligence data and enemy information reports" between military officials.

The New York Times and The Washington Post have recently reported that Maven relies on Anthropic's AI technology. Since 2022, Palantir has also sold another intelligence platform to the U.S. Army—the Army Intelligence Data Platform (AIDP). This platform integrates data from Maven and at least four other government systems, supporting pre-operation intelligence preparation, graphical presentation of troop and weapon locations, and features a tool called "Dossier" for generating continuous battlefield intelligence assessments. It is unclear whether Claude has been integrated into AIDP.

AIP: The AI Interface Layer Embedded in Operational Systems

Palantir integrated Claude into military systems specifically through its Artificial Intelligence Platform (AIP). AIP is not an independent platform; it is an application layer running on existing Palantir commercial products (such as Foundry or Gotham), providing users with a chatbot interface for queries and task execution. Palantir refers to this as the "AIP assistant" or "AIP agent."

The AIP assistant is powered by third-party large language models from Anthropic, Google, Meta, and others. Customers can choose which model to use, and which training data the model accesses to generate responses. This structure is especially meaningful in intelligence and national security contexts—classified intelligence data can be restricted as exclusive data sources for the models.

A demonstration video released by Palantir in 2023 showed how the AIP assistant helped a "military operator monitoring activities in Eastern Europe" plan and initiate a ground attack on multiple tanks. The entire process was conducted via dialogue with the chatbot: the system first sent an automatic alert about "potential anomalous enemy activity," then the analyst requested MQ-9 Reaper drones for reconnaissance, and instructed the AIP assistant to "generate three action plans to strike enemy equipment." The assistant provided three options within seconds: aerial assets, long-range artillery, or a tactical squad.

Subsequently, the analyst asked the assistant to "analyze the battlefield," "generate routes" for troops to reach enemy positions, and "deploy jammers" to disrupt enemy communications. After final review, the analyst ordered the troops to move out. In this scenario, Claude serves as the AIP assistant's "language layer," interpreting commands and generating responses.

Another NATO-related demonstration displayed similar logic: analysts viewed troop and weapon deployments on a digital map, and, after clicking a button, a tool powered by GPT-4.1 generated five possible military strategies, one titled "Support-Fire, Penetrate, Shock and Destroy." The demo also showed analysts could select different AI models within the interface, with Claude listed alongside ChatGPT and Meta’s Llama as options.

Claude’s Role: Intelligence Generation and Analysis

In addition to real-time combat assistance, Claude is used to generate intelligence assessment reports. In June 2025, Anthropic’s Head of Public Sector Kunaal Sharma conducted a demonstration showing how Claude's enterprise edition could generate an "advanced" analysis report for Ukraine's drone strike operation "Operation Cobweb."

In the demo, Sharma prompted Claude to create an "interactive dashboard" with operational information, convert it into "object types" for analysis within Palantir’s Foundry platform, write a detailed analysis of recent developments in Russian border provinces, and draft a 200-word summary of the "military and political impact" of the operation. Sharma noted such reports typically take hours to create manually, whereas Claude can generate them in very little time. He added that through cooperation with Palantir, the federal government can access internal datasets beyond public information.

When Palantir announced its partnership with Anthropic in the military and intelligence fields in November 2024, it stated that Claude’s integration helps analysts discover "data-driven insights," identify patterns, and support decision-making in "time-sensitive situations."

The exposure of these systems comes against the backdrop of rapidly deteriorating relations between Anthropic and the Pentagon. In late February this year, Anthropic refused to grant the government unconditional access to the Claude model, insisting that its system should not be used for mass surveillance of U.S. citizens or fully autonomous weapons. The Pentagon subsequently classified Anthropic’s products as a "supply chain risk," prompting Anthropic to file two lawsuits this week, alleging this action by the Trump administration constitutes unlawful retaliation and seeking to overturn the determination.

This dispute is refocusing public attention on a central issue: When AI model developers and the military disagree on the boundaries of use, how will deeply embedded AI technology in combat systems be constrained? Reports indicate that Claude is still used in certain U.S. defense operations, including those related to the Iran conflict.

Risk Disclaimer and Liability ClauseThe market carries risks, and investments require caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their specific circumstances. Investing accordingly is at your own risk. ```