The Department of Defense has inked formal agreements with Nvidia, Microsoft, Amazon Web Services, and Reflection AI to deploy their AI systems on Impact Level 6 (IL6) and IL7 classified networks — the highest security tiers covering systems critical to national security. Over 1.3 million DoD personnel already access GenAI.mil, the Pentagon's secure generative AI platform. The deals were explicitly structured to prevent "AI vendor lock-in" by drawing on a "diverse suite" of American AI capabilities.
The context is critical: these deals fill the vacuum left by Anthropic's refusal to grant "unrestricted use" of Claude/Mythos without guardrails against domestic mass surveillance and autonomous weapons. Anthropic's principled stand cost it the classified contract but handed Nvidia, Microsoft, and AWS a direct path into Pentagon AI infrastructure. The irony — Anthropic's Mythos is simultaneously being used by the NSA for vulnerability detection while the Pentagon routes classified AI through its competitors — underscores how fragmented US government AI strategy has become ahead of Anthropic's expected IPO.
TechCrunch · May 1$app:) without
requesting extra merchant permissions in Admin API 2026-04+; reduces merchant friction during
app installs and upgrades.
[link]