The only reason I would say this would make any sense would be to have the ultimate offsite backup, but even then would probably be cheaper just to have 3 or 4 of the same hardware spread on earth.
A few things they seem to have not considered:
- 1st, and most important, cost of launch, even with starship costs of launch all the hardware is REALLY heavy, it could still be >10x the cost of the hardware itself
- 2nd, power generation, solar has come a long way and in space it is much more efficient by area, but the amount of solar panels that would be needed to power even a modest datacenter is insane, and how to deploy it is not trivial by any means, and it can get to the point that the size creates quite significant stray current only by traveling through earths magnetic field, I remember a couple of decades ago an experiment that produced electricity by extending a big wire in space and the voltage generated was high enough to break the insulation layer and damage the base which snapped off the cable
- 3rd, heat dissipation in space is a b*tch, people think is cold, but forget that it is just a vacuum, which is a VERY good insulator, half the time you are being heated by the sun (if you don't choose a sun synchronous orbit which would be all the time) and there is no medium to dump that heat, you can only radiate it away, which is hard, and based on the ISS issues, also non-trivial and when a problem happen you are literally toast
- 4th connection bandwidth and latency, even with the best laser connections we are talking a few dozen gigabits at best per channel, compared to terabits per fiber on earth, and depending on your orbit you can have "low" latency but need to reroute the connection constantly with a LOT of jitter or have a "fixed" connection with insane latency, both have big downsides unless you are doing simulations or AI training only using local data and sending back only the final result, anything else the latency would be annoying to the end user
- 5th impacts and kessler syndrome, with the necessary solar panel and radiators area this would be a HUGE target for mircro meteorites and orbit garbage impacts, basically impossible to steer out of the way and just create more and more small objects with time
- 6th de-orbiting, how would this would de-orbit? ISS would be a object in comparison and even it is complicated to get down safely, imagine something that is easier to be measured in acres instead of square meters/feet, and the amount of metals in the atmosphere could be significant to the point of causing issues to the high atmosphere
I could probably go on, but I think it would be much easier and cheaper going the microsoft route and just put the datacenter under water in a cold area of the world.