It is not failure rate, it is the cost if it happens. I work for large corporations and the calculations are entirely different from what a small, cash strappe

It is not failure rate, it is the cost if it happens. I work for large corporations and the calculations are entirely different from what a small, cash strapped startup would be doing. For example, if you work on a new product and it would take 1 day to get Mac Mini for free vs you would be paying for cloud solution 1000 dollars daily but get it one day sooner, it might still be well worthwhile to get cloud version. Because if you think about it, it is possible that Mac Mini would be holding up your product development for one day, then you calculate how much you hope to earn from that product daily and how much you are loosing while you haven't launched and then it is no brainer. As I said, it depends on business case. For some, the cost of this M1 VM is not going to be any factor in the decision. Have you seen how expensive GPUs are in the cloud? Have you seen how many clients are that pay for huge farms of these?