Anthropic has been pretty transparent that their plan for dealing with Chinese open models is regulatory capture. All the fear-mongering and posturing that AI is so dangerous only they can be trusted with it. They want to legislate their competition away, at least in a few markets. OpenAI has been more subtle about it.

It's a dangerous gamble, and bad for the US. But, given how cheaply many of our politicians can be purchased and how much money is flowing into the leading AI companies (and a bunch of also-rans), I guess we should expect to see them rolling the dice on that gamble.