My impression is most commercial insurance policies specifically exclude acts of war (and similar). The extent of systematic damage that can occur in such scenarios would easily put them into insolvency. There are exceptions for life insurance, maritime shipping, aviation etc. but I gather they're uncommon, often limited-time in nature and come with high premiums. I've never heard of the equivalent for buildings and contents. Maybe someone can enlighten me.

I also agree with the sibling comment that suggests even if there were outsourced options available, hyperscalers might do better to self-insure.

But regardless whether the company eats it, or pays it through insurance premiums, I'm still curious how this type of risk is planned for (to whatever degree it is) and accounted. I assume it must have been contemplated in a deliberate manner, somewhere on a scale of "We knew this could happen, considered x, y and z, decided the venture was justified, and here's how we planned for such a contingency" to "That was dumb siting and someone will be fired". Obviously anywhere you put a building entails some level of local risk.

As years as a volunteer firefighter we do a lot of risk management at the tactical level and have to think through assessment and potential consequences. There are lots of guidelines we learn to help produce sound decisions. But at the end of the day to apply them you may need to make a judgement call, and you need to be prepared for things to go wrong. In a way I guess I'm asking about the economic and operational equivalent at the scale of hundred-million dollar data centers.

https://xkcd.com/1737/

Ps. Like the original commenter said, I'm in no way meaning to be insensitive to the larger human and regional consequences.