How) should data be incrementally reclaimed from the cloud?

Company A exclusively utilizes a small set of domain-specific SaaS products to store literally everything. There is nothing of importance on local machines. They are essentially thin clients for a small set of cloud services.

This strikes me as precarious, given that everything breaks the moment that something happens to the local internet connection or service providers' connection (the latter, but not the former, being highly unlikely). It's turtles all the way down, and if a single turtle slips and falls so does the business at the top. I'm the type to use plain text formats stored on a local server, use WAN as little as possible, and keep a gun next to the printer if it makes any weird noises[0], so I'm (opionatedly) suspicious of 'too much' cloud.

The thing is, Company A's setup works. Functionally, the amount of abstraction they're playing with is quite powerful, and it makes everything pretty easy for them.

The following questions are asked from a technical perspective, not an inter-personal/workplace perspective:

- Should I work to keep this company's data a bit more independent? If everything is working just fine right now, is it really best to start messing with things?

- If so, how do I do this incrementally?

0: https://news.ycombinator.com/item?id=20981980