> If you’re going to go multi-region and take the latency hit may as well go multi-provider no?

No. If you go multi-region, you use the same tooling, same terraform modules and logic and so on. There's little plumbing needed to make it work. And latency wise this is not an issue in most cases, since most of the requests are covered by the CDN anyway. And you don't have to duplicate everything.

If you go multi-cloud you need to learn a whole new set of systems. And that is expensive. Both in terms of operating and people - because you will need more.

> Agreed, but thats why all those who justify the 10-90x premium of the cloud over bare-metal are full of it - that premium is not actually worth it.

You get charged a premium for convenience. And a high enough chance you don't have downtime.

> Maybe for you it’s not a secret? Literally every thread tries to justify cloud reliability and their resulting markups.

Cloud is reliable if you are willing to spend some money to benefit from that reliability and convenience.

---

Besides this, another thing where cloud saves you money is compliance. They have all the right attestations in place to make your audits go easy. If you self-host on bare metal, you're going to spend a lot of time to be compliant with various regulations. Maybe if you're a small company, you don't have that much compliance you need to do. But once you grow a little, those immutable Stackdriver logs are a godsend when you're asked to prove logs have not been tampered with.