You're right. It's still not as good quality as the pile of HPE and Cisco kit we replaced with it and it's more expensive at the TCO level.

This is assuming you have a competent infrastructure and networking team, which is expensive. Lots of middle managers don't like hiring engineers that earn more than they do. It defies the pecking order.

If you go all-in on cloud, you now need a team of cloud experts and you are beholden to whatever pricing strategy AWS/GCP/Azure deem appropriate in future.

It's also a terrible deal for the engineers who have to learn all this cloud stuff - instead of learning general computing, you learn how to work with vendor X. If they change things, go out of business, or get too expensive, that knowledge becomes useless.

If computing is a core part of your business, you should get good at it.

We have many more people running our cloud that we had infra. It's far more complex.

Easier to hire lots of low paid people than a few highly paid engineers. I don't dispute that TCO including labor is generally higher with cloud infrastructure.

All our engineers cost more than the DC engineers we had before.

If aws is the top quality cloud provider, then might as well give up now and buy a server a rack and an ac

good luck getting the same type of offering from vmware

Kubeadm? Ansible?