The problem is that, technically, an IP address doesn't have a country. GeoIP is at best a suggestion, and it is trivial to route your traffic through a "foreign" IP.

It is about as enforceable as "don't read letters written by redheaded people": it'll never be airtight, and it'll always have a huge amount of false positives.

The only possible "solution" is to ban all open devices from the internet, mandate a DRM-style "trust zone", have the device do mTLS, and embed a realtime DNA sequencer in every device.

Yes but if you don’t have any assets in the country you could not liable, that was OP’s point.

But even being based in Switzerland I believe there’s some inter-countries agreements, so it would not be practical.

They wouldn't be providing services in France, they'd be providing service in Switzerland to people from France. No different than if someone from France traveled to a bagel shop in Switzerland for a bite to eat. If France wants to make it illegal for French citizens to communicate with a foreign business in Switzerland, I don't see how it's the business's responsibility to enforce that.

As a side project I help out with some technical stuff at a hosting company that provides services to a political group that is strongly anti Erdogan. The infrastructure is in the USA.

We do not block IP traffic from known Turkish ISP ranges. The customer (the political group) receives all sorts of dire legal warnings and threats that what they publish is contrary to Turkish law.

None of our other hosting platforms are configured to block any traffic to/from Turkish IP ranges. We have no business presence or financial involvement with anything domestically in Turkey. Am I subject to Turkish jurisdiction? I don't think so.

But by what you wrote, I am providing services in Turkey.

Hence the “shield” part. I guess it’s easier said than done when you’re a Swiss non-profit and not an anonymous sysadmin in Iraq (or whatever jurisdiction would work for this example, idk).