How US Tech Giants Are Losing Their Independence

The US Department of Defense and Anthropic have spectacularly failed to reach a $200 million deal. Consequently, the Pentagon promptly classified the AI startup as a security risk after negotiations regarding the use of Claude models for military purposes collapsed. Meanwhile, rival OpenAI wasted no time in securing the government contract. This unprecedented event marks a new era of state intervention, where Washington dictates a simple rule to private tech firms: National security trumps corporate ethics.

The fallout is the result of weeks of tension and personal friction between the negotiating parties. At its core was Anthropic CEO Dario Amodei’s refusal to grant the military access to his AI model for mass surveillance of US citizens and the operation of fully autonomous weapon systems. While Amodei insists on ethical guardrails, the Pentagon demands total operational freedom. The escalation culminated in a hard deadline and Anthropic being branded a "supply chain risk"—a label previously reserved almost exclusively for foreign threats.

Where Anthropic hesitates, Sam Altman accelerates. Almost simultaneously with the competitor's exit, the OpenAI CEO announced an agreement to provide its technologies for classified military systems. OpenAI accepted the demand that its AI be used for all "lawful" purposes, provided that technical guardrails broadly uphold its own safety principles. Within the industry, this move has drawn heavy criticism, with many accusing OpenAI of sacrificing ethical standards for a strategic advantage over Anthropic.

The consequences for Anthropic are drastic. Labeling a US company as a supply chain risk is a novelty that jeopardizes not only future government contracts but also the trust of global investors. Due to this classification, any company holding government contracts is now prohibited from working with Anthropic. This case serves as a precedent for whether companies can effectively limit the use of their technology by state actors. Anthropic is reportedly preparing a lawsuit against the risk-factor designation in an attempt to save multi-million dollar contracts.

Reality is outstripping politics at record speed. Despite the ban, the US military reportedly used Anthropic’s Claude AI as recently as last weekend for airstrikes in Iran. The system was utilized for intelligence analysis and target identification, highlighting the deep integration of the technology. The now-mandated six-month phase-out forces agencies into a risky and complex transition to competing products like ChatGPT while operations continue.

This conflict heralds a new phase of massive state interference in Silicon Valley. The message from Washington is clear: those who do not comply will be politically and economically isolated. When the US government begins actively shaping the tech stacks of private companies, it changes the rules of competition and sovereignty across the global AI market. Even conservative voices are warning against such intervention in the private sector. Dean Ball, a former Trump advisor on AI, describes the situation as "corporate murder" and currently advises against investing in the US AI market.