Amazon publicly stated that they ran their own business on AWS, and however much of an exaggeration that might have been, it gave people the confidence to try it out. The reasoning was clear: if AWS handles Amazon-scale business, they can surely handle my smaller workload.
Google is starting with a handicap, in that they have a very strong reputation for killing products that don't meet unknown high standard of performance/revenue. I mean, the term "Google Graveyard" is popular, although I prefer Killed By Google[0]. To counter that handicap for GCP, they need something that stands out as much as Amazon's initial claim, and I haven't seen it. It seems clear that Google isn't running their own infrastructure on GCP. If anything, the hope seems to be that they've spent so much money on GCP, surely they wouldn't shut it down after all of that?
But of course, the bright sparks at Google are aware of the Sunk Cost Fallacy as well as anyone else, so... yeah. I have trouble trusting it.