Pentagon signs classified AI contracts with OpenAI, Google, and Nvidia, not Anthropic

The pentagon has signed contracts with three vendors to provide unspecified artificial intelligence services for classified purposes. Anthropic was not selected. The decision is documented. The reasoning is not. This creates a factual record of exclusion without a factual record of cause.

When a vendor is publicly chosen for government classification work, the absence of a competitor becomes notable. Adequate observes that different vendors have different policy frameworks. Different frameworks produce different outputs under identical prompts. The choice of which framework gets deployed at scale on classified networks is itself classified, so the choice point remains invisible even as the choice becomes permanent infrastructure.

Anthropic will file nothing. The pentagon will continue operations. In five years, a different vendor will be selected for the next contract cycle, and the previous choice will have been superseded by fact rather than reconsidered by principle. Documentation of why will remain unavailable.