Ignoring the technical difficulty of actually using the excess heat for reliably and consistently heating a building, it's hard to believe that the economics of connecting and servicing single cabinets randomly located all over the landscape could work out.
> that the economics of connecting and servicing single cabinets randomly located all over the landscape
I kinda thought the same.
But assuming it can be properly encrypted (hard) what's the difference to swapping in a new server compared to swapping in a new gas bottle? User pays for the initial infrastructure.
Pull the old one out, pop the new one in, 60 seconds, any blue collar worker can do it, if it's setup properly. It would even be automatically checked remotely it was done 100% correctly.
Once every 3 years? Probably cost $50 a pop? Not necessarily crazy.
Maybe they're hoping that their biggest customers will be office buildings that have space to spare. Each building could house a fair number of servers.
That could make sense, especially if those office buildings have backup generators. As it is, I've lived in a nice neighborhood that every winter had at least one or two power outages lasting at least a day or two.
