Google, Microsoft, and xAI will allow the US government to review their new AI models

The notable absentees matter as much as the signatories. Anthropic and Meta are not part of this framework, and given Anthropic's recent friction with the Pentagon over classified contracts (covered here in early May), its absence from a Commerce Department review program is a second consecutive signal that its relationship with federal institutions is under strain.

This connects directly to the Pentagon AI deals story from May 1st, where Google, Microsoft, and xAI all secured classified defense contracts while Anthropic was excluded. That pattern now repeats in a civilian regulatory context, suggesting a consolidating tier of 'government-aligned' labs that are willing to accept pre-deployment review as a cost of access. The ethical divergence benchmark covered by The Decoder on May 3rd adds a second layer: if frontier models encode meaningfully different value systems, a voluntary government review process becomes the de facto arbiter of whose ethics get normalized at scale, which is a structural advantage for labs inside the framework.

Watch whether Anthropic formally joins or publicly declines this Commerce Department framework within the next 60 days. A refusal or silence would confirm a deliberate positioning strategy around independence from federal oversight, which would sharpen the competitive divide already visible in the Pentagon procurement data.