>they can only launch things that are capable of serving billions of requests per day

They did not have to launch something like ChatGPT at full scale. I understand very well the cost associated with it. But, they should have been the one setting the stage and the narrative. Building expectation releasing cool demo, experimental features backed into Docs. Things that could show where "potentially" Google could have taken the technology in time ..

But, it seems like, that while they were working on the "perfect working money printing AI product", they completely left the stage to OpenAI and friends.

I think, that was definitely a strategic error.

Perhaps Google will come ahead, who knows. I hope this was a wakeup call for them.