[Jugurtha]

>It’s an old and lazy/annoying complaint.

Granted, you offered a nuanced reply for someone building something on these clouds. However, when you're dealing with enterprise that are directly competing with the main cloud providers on other verticals, they are not being lazy or annoying, they are being careful, and not without a reason, and avoid using their competitor's cloud infrastructure.

Some of our clients have no problems using these cloud providers. Others wouldn't go through them because that would leak information.

If I'm not mistaken, Google used to buy datacenters in stealth mode, under different companies created for that purpose, to avoid getting on Microsoft's radar and keeping how successful search was from them.

We use GCP, for context.

[ignoramous]

> Google used to buy datacenters in stealth mode, under different companies created for that purpose, to avoid getting on Microsoft's radar and keeping how successful search was from them.

I believe, Google for a long time (and still does?) thought their infrastructure was the secret sauce, and that might have impeded them from competing with AWS in the early years (despite having all the pieces in place already)?

[Jugurtha]

There are apprehensions like these all over the place to protect their user base.

This matters to us because we're building our machine learning platform[0] and we want flexibility as opposed to using their machine learning products, because each considers themselves "the only cloud". Therefore, we're compelled to do "multi-cloud"[sigh], because we want to be able to train models on X, deploy on Y, and have data on Z.

It's funny to watch, though, as if I recall correctly, there was an article on one of these companies where saying "multi-cloud" was blasphematory.

- [0]: https://iko.ai