Earlier quoted context omitted.
You don't need hundreds of services. Give me virtual machines, reliable block storage, file storage and object storage, networking, dns, managed kubernetes, and it will cover the majority of workloads in Europe that run on Openshift or Openstack today.
> Give me virtual machines, reliable block storage, file storage and object storage, networking, dns, managed kubernetes, [...] But managers wants to _buy_ these services, not be directly responsible for them. That's where the problem lies, as I see it.
I don't know what managers have been reading/hearing, but for the last decade or so as a developer what I've mostly been hearing is that the only people who actually benefit from Big Data architectures are FAANG, that it's much cheaper to run on a single small self-hosted system that's done right, that the complexity of managing the cloud is even higher than a local solution.
This matches my own experience of what people needed to serve millions of users 20 years ago. If you can't handle a chat system or a simple sales system with 100k-1M customers on a server made out of one single modern mobile phone, you're either just not trying hard enough or have too many layers of abstraction between business logic and bare metal. Even for something a bit more challenging than that, you should still be thinking thousands of users on a phone and 10k-100k on a single device that's actually meant to work as a server.