Theory: Certain USG employees are going after Anthropic because they (or someone they know) has a financial stake in OpenAI. OpenAI has made the same claims, and months ago released "dangerous" security-analyzing models which "need limits", but USG never punished them for it.

Additional theory: Altman is behind it.

That's a huge grasp. Anthropic have been making this bed for years now. Altman did not need to do a single thing for this outcome to materialize.