The American army is taking a strategic step forward by massively integrating artificial intelligence into the heart of its classified operations. The Pentagon has just selected seven technological giants to power its military capabilities with AI, while excluding a recognized player. This choice, far from being trivial, reveals both a desire for technological acceleration and deep tensions between defense imperatives and ethical red lines within the industry.
In brief
- The Pentagon formalizes partnerships with seven technology giants to integrate artificial intelligence into its classified operations.
- These AI models are intended to improve data analysis and assist military personnel in complex environments.
- Despite its recognized performance, Anthropic is excluded from the agreements due to a dispute with the American government.
- This decision reveals growing tensions between military requirements and ethical limits defended by certain companies.
A strategic alliance between the Pentagon and AI giants
While Washington is still seeking an agreement with Anthropic, the US Department of Defense has formalized the signing of agreements with seven major technology companies to access their artificial intelligence models for classified operations.
The Pentagon precise that these tools will be mobilized for critical missions, with a stated objective: “These partnerships will accelerate the transformation of the US military into an AI-focused response force”. The models will be deployed in level 6 and 7 operations, the highest in the American military hierarchy. Here are the companies chosen:
- OpenAI;
- Google ;
- Nvidia;
- Microsoft;
- Amazon via AWS;
- SpaceX (via xAI);
- Reflection.
In detail, these systems must “make data synthesis, context understanding and […] contribute to a fighter’s decision-making in complex environments”.
Artificial intelligence therefore provides analytical support in situations of high operational intensity. The ministry insists on one point: critical decisions, such as choosing targets or launching strikes, remain under human control.
Anthropic sidelined and ethical tensions within the ecosystem
At the same time, the Pentagon chose to exclude Anthropic, despite the recognition of its Claude model as one of the most efficient. This exclusion is part of an open conflict with the American administration, which decided at the end of February to terminate all its contracts with the start-up, a decision contested in court. The dispute originated in Anthropic's refusal to authorize the use of its technologies for applications related to mass surveillance or lethal operations.
According to some on-chain data, this divergence illustrates a significant divide within technology companies. More than 600 Google employees signed a letter asking their management to stop providing AI models for classified operations. Faced with these tensions, the Pentagon is adopting a clear strategy: multiply suppliers in order to “to avoid being dependent on a service provider and ensure long-term flexibility”.
The ongoing development goes beyond the simple technological framework. It raises the question of the role of private companies in defense infrastructures and the governance of high-criticality artificial intelligence systems such as Claude AI. As these tools become integrated into military decision chains, trade-offs between performance, human control and ethical acceptability could become a major point of friction, with lasting repercussions across the entire technology ecosystem.
Maximize your Tremplin.io experience with our 'Read to Earn' program! For every article you read, earn points and access exclusive rewards. Sign up now and start earning benefits.
