Just found this thread... actually, moving European companies (at least most of them) on European cloud infrastructure isn't that difficult. All it takes is to use the current tariff madness to introduce substantial tariffs on cloud services provided by overseas companies in the EU. It needs a good definition of "overseas company" as Internet and borders... not very clear most of the time.
I kind of wonder whether that's true in a practical sense. Well, in a narrow sense it is because in theory one can always make up a big enough penalty, but at some point it becomes politically and economically unviable (even more so than any have any tariff at all, I mean), and if too little I suspect a lot of companies would swallow the costs if cloud providers pass them on to their EU customers, esp. if cost of migration would be higher than their next quarterly report. (NB: I am not an IT person.)
Specifically, what I'm curious about are effects of large global presences and and vendor lock-ins. Say you have some EuroCom that mostly does business in the EU but has some customers South America, and its cloud services are hosted by Amazon (which IIRC is the biggest global player and mostly caters to larger businesses), and not many cloud services have a presence there. So because Brazil has some laws about hosting their citizens' data, an alternative cloud provider needs to have a presence in São Paulo or wherever, and if the tariff is big enough to make EuroCom seriously think about getting rid of their business there, your tariff is arguably hurting the EU more than it helps. Meanwhile, Amazon does a
lot of things, and does them in a peculiar ways dependent on their own APIs and infrastructure. Say EuroCom has a bunch of geographically separated servers that mount the same disk, or a serverless database that's automatically replicated across many locations globally; none of them are actually irreplaceable, mystical features, but it does mean that EuroCom would need to think very carefully about how to untangle them for migration and possibly permanently get more tech monkeys for their NFSs and more DBAs for their databases, which is potentially yet more costs. Meanwhile, they also probably have a contract with Amazon for the next several years anyway, so it's easier to raise a political fuss about internet tariffs being bad than seriously consider migration.
I don't know... little cloud providers like Linode and Digital Ocean are one thing, but how different is that for the big fish?
(Naturally, I don't want to imply that any of that is in anything close to truly insurmountable, but I have doubts it's as simple a problem as that, and in particular I suspect it may require investment and/or more complicated economic incentives.)