Well that is kind of a problem of their own making. The clouds refuse to entertain the prospect of pre-paying for services/having some sort of hard spending limits because they know that over-allocation is probably driving a decent amount of revenue.
That’s not how quotas work in GCP. Google sets quotas for certain APIs for interacting with GCP itself, like how many VMs you can create per second. They’re not billable. Sometimes these quotas can be be increased if you need them to be. But the way op described it makes no sense.