Anthropic's $25B AWS deal quietly unbans Claude CLI agents

openclaw claims claude cli is “allowed again” for agent-style usage (not confirmed publicly by anthropic)

at the same time, anthropic just deepened its deal with amazon – $5b invested now, with up to another $20b committed over time to trainium and aws infrastructure

the timing is hard to ignore:

- running agents was “misuse” when it stressed infra and pricing models
- now there’s effectively unlimited compute backing one of the main model providers

when compute is constrained, usage gets restricted. when compute is funded, usage gets normalized.