Pentagon Classified AI Training Enclaves Expanding Despite Anthropic Ban

The Pentagon is expanding classified AI training enclaves across IC agencies even as the Anthropic ban forces a rapid migration away from Claude-based systems. Multiple agencies had integrated Claude into classified workflows under the July 2025 contract, and the six-month phase-out is creating operational disruption. OpenAI and Palantir are positioned as primary replacements, though neither has completed the full security accreditation process that Anthropic had already cleared.