Grok has plenty of "non-woke AI" cred which will make it a preferred choice for lots of government-side and politically sensitive workloads.
Anecdotally, Grok is mostly not that anti-woke. It is coasting off of its reputation from Elon's marketing. That said, it does have meaningfully fewer guardrails, which is a real benefit.
Nazi AI is bad actually, most governments will not want to be associated with it
I am pretty sure the gov gets an uncensored /heretic version of the AI models instead of the nanny stuff we get so Grok really doesn’t have a “killer” feature. Even for multi model passes(I.e used purely as a contrarian gate) it’s not really that great due the high rate of hallucinations.