Sabotage works by introducing friction into your opponents activities. Sabotaging one piece of one data center doesn't do much, but the more you do, the more outsized the impact.

Imagine I'm a factory building widgets. If I buy materials, my default assumption is that I get the materials I asked for. If 5% of the time, or even 1% of the time, my vendor sends me junk that breaks my machines, now I have to introduce a step to verify that the vendor sent me the right ingredients to every widget. That's an asymmetric cost.

The messaging for something like this wants to be "we publicly announced and took credit for this this time", because it's good publicity, and the threat of future, clandestine attacks increases costs across the board. If you can include exactly how you did it, you might even inspire copycats.

This is also all the sabotage the saboteurs have volunteered to tell you about. If your opsec has allowed sabotage to happen, it’s prudent to assume there’s other sabotage you don’t know about.

> my vendor sends me junk

Indeed. A single bad review of a product from a user, if justified, can build the impetus to destroy a product. Three bad reviews probably will.

Insurance costs too can be affected.

That's right, which is why saboteurs need to be dealt with on the harshest terms.