I think this particular scenario is far fetched as that would be economic suicide for the US, an empire-ending decision. And while not everyone has backups, many/most of the important companies do so they would eventually recover.

> Many European companies would stop to a halt as they can't access any documents they have "on the cloud" or maybe can't even access their own phone or computer.

I hate that "Nobody got fired for choosing IBM" is a thing and that the people suggesting that we have good enough FOSS options when things were being planned out were probably given a dismissive look by the business people who were promised the sky by MS salesmen.

At least that's how I imagine it probably looked, given my own past experience of suggesting PostgreSQL and in the end the project going with Oracle (it's okay when it works, but for those particular projects PostgreSQL would have worked better, given the issues I've seen in the following years). It's the same non-utilitarian / cargo-cult thinking that leads to other solutions like SQLite not being picked when the workload would actually better be suited for it than a "serious" RDBMS with a network in the middle.

Apply the same to server OSes (Windows vs Linux distros and even DEB based distros vs RPM RHEL-compatibles), MS Office vs LibreOffice when you don't even need advanced features and stuff like Slack/Teams vs self-hosting Mattermost or Zulip or whatever. It's not even jumping on untested software, but fairly boring and okay packages (with their limitations known that are objectively often NOT dealbreakers) and not making yourself vendor-locked (hostage).

I guess I could also make the more realpolitik take - use MS, use Oracle, use whatever is the path of least resistance BUT ONLY if you're not making yourself 100% reliant on it. If Microsoft or Google decides they hate you tomorrow, you should still have a business continuity plan. If systems have standby nodes, why not have a basic alternative standby system, or the ability to stand up a Nextcloud instance when needed for example (or the knowledge and training on how to do that)? If people had govt. services before computers being widespread and you can have people processing a bunch of paper forms, then surely if push comes to shove it'd be possible to standup a basic replacement for whatever gets borked while ignoring all of the accidental complexity (even if it'd mean e-mailing PDFs for a while). Unless someone builds their national tax system or ID system on a foreign cloud, then they are absolutely fucked.

I don't think it's easy to replace ENTRA feature-wise with European provider.

Or github if you're using a bit more than self-hosted gitlab can provide.

It's not always about the location, it's usually about features (how it integrates into other hardware/software) rarely prices.

For example, can you suggest firewalls for offices that aren't either American or Israeli? We'd need something to replace Palo Alto, Bluecoat, Fortigate and Juniper. Also it'd be good to replace Cisco VPNs to be honest.

But it kind of must be feature parity, because (European) regulators hold our balls over hot coals.

Sophos