It wouldn't be that hard for them to pull off. I'd actually be shocked if they wouldn't put assets in the area for exactly that thing. It would be a hindrance to all the defense firms that are indirectly supporting the US military. It would also be a massive annoyance to both corporate interests and voter interests domestically.
When what brought it about, speaking in a hypothetical future tense, was US involvement that benefits no person or company in the US it would quickly deflate the US's political will or mandate from the people to pursue these wars. The current mandate that they have is that the current party technically controls the military, and so far it hasn't negatively impacted the public. The second that's not true, the American public will be asking questions about how we calculate who to attack. And there will be no answers besides open admission of corruption.
It would be very smart for Iran to do. It may even improve Iran's popularity in the US because people don't really like those data centers. And double smart when no one likes Israel at all in the US right now. How ironic will it be when Iran's PR in the US is better than Israel's, and we are fighting a war that only benefits Israel? It would divide the people from their government even more and knock out a strategic war resource.
But to protect PR, they'd have to somehow do it while minimizing deaths. These buildings have almost no personnel in them. But if there is one guard in the whole building, that will end up being the story. The news in the US that works for the military in times of war knows that US citizens don't give a single fuck about an AWS blade server. But one single AWS guard might be enough to get the American public to actually care about the Iran issue. So that is the risk they have to balance with that kind of an attack.
They don't need intercontinental missiles to do it. They only need rockets more crude than the ones they have given to Hamas or Hezbollah to target something 100 yards in front of them.
Hahahaja
Twenty years ago I toured a data center not far from DC.
It was specifically designed to prevent bomb attacks.
The data center itself (multiple buildings) was set back from a blast wall that was higher than the data center buildings, and several feet thick.
The perimeter of the property was landscaped in a way to prevent vehicles of any type from simply charging in.
They brought power in from 4 directions, from 4 different, independent systems.
They could run for 2 weeks on generator power - they had at least 2x the generators required to run at 100% capacity, so they could maintain one generator while the other maintained load - multiple generators for each building, with enough fuel on site (also protected by being deep underground).
These were all things they pointed out and were pretty obvious. Now think about all the things they didn't tell us about.
Oh, and at the time they had a duplicate data center in Florida which had massive data lines connecting them. THAT data center could withstand a direct Cat5 hurricane, including contending with associated flood levels.
There are thousands of DC's in the US, and I'm sure people who deal with critical data have failover plans that wouldn't even fart at your ideas - I've been on teams making such plans, and for just critical business systems you'd could take out all but one DC where we housed crucial systems and we would only know because we'd get alarms, not a loss of critical functionality.
I can say this because I've dealt with it many times (and I'm just a peon in these businesses, one of many who provide expertise for such things).
They wouldn't put that much effort into it if it weren't a legitimate threat that could occur from an anticipatable motive paired with an anticipatable capability.
The fact that they have designed the buildings that way proves all the points I made.
A threat they've mitigated
You're funny, acting like you know better than the teams of people who've worked on mitigating multiple aspects of threat modeling.
As I said - they showed us what couldn't be hidden, or were permitted to show us.
There's a few other elements they talked about which I won't repeat here, but it really drove home how much time they'd put into understanding worst-case.
About the only way you'd damage these places is with a direct bomb from overhead - and even that would take a significant direct hit.
Remember, this is one of several for this data center company. How many others are there that live replicate the important systems?
For business I've worked with systems which we replicated (live, transparent auto-failover) across five data centers in different zones. These zones are defined by their risk profile - both natural (hurricanes, floods, tornado, etc) and human threats such that no single event could disable them all.
These "normal" data centers (probably not as hardened as the DC one) also maintain multiple power systems, and redundant data connections via independent routes.
This for just business-critical systems.
What I'm saying is, you are utterly talking out of your ass about something you have no fucking clue about, and your hubris is tiring.
Again, I'm a relative peon to this stuff, and can see how much thought's been put into security.
Who designed it?
Asking for a friend, AWS Dubai.
I honestly have no idea (its been 20 years). I can't even remember the name of the place.
Lol, I'm pretty sure they're not too worried - redundancy is the name of the game. And they planned their risk-offloading long before they even broke ground.
So while it may appear Amazon would lose something, they won't, as their contracts and insurances assure the risk is transferred to someone else. Its how large business operates, and part of why cloud exists at all - move CapEx to OpEx while gaining the ability to blame your cloud provider for failures and maximize flexibility.
As an IT analyst type I hate it but can't argue against it from a cost/benefit perspective.