Comment Re:Should have hired from Meta or Google (Score 1) 10
If Snowden taught us anything at all it's that the NSA are light years ahead of what people think is possible as far as data collection, storage and processing is concerned. Only relatively recently it was reported that the US were buying huge amounts of commercial information off the open market. I can't believe for one second they're not also routinely augmenting that with OSINT data on top of that. In fact it would be incredibly naive to think that they wouldn't be doing so on a mass scale for decades at this point.
Google and Meta will also be incredibly good at dealing with masses of data of course, but their focus is very different and I'd suggest the volumes involved, as huge as they are, are still smaller than the NSA. But I don't need to speculate. Back in 2013 the NSA released a letter that revealed it was "touching" about 30PB's of data a day which at the time was 50% more than Google at 20PB. It is a reasonable assumption that they've at LEAST kept pace with that. I think it's a pretty safe assumption that they are handling mutiple exabytes of data a day at this point.
There are four massive datacentres dedicated to this work that are known. The one in Utah cost 1.5 billion to build and the governer of that state is on record saying it was designed to hold a yottabyte of data. That's roughly 45 million 22TB HD's so it's not beyond the realms of reality when there's plenty of companies with exabyte storage.
xKeyscore is already known to handle petabytes of data even 10 years ago so by now it's not unreasonable to assume it's handling 10's or even 100's of exabytes. When you have 10's billions available to you "for free" and you have a singular goal with no sharehilders to worry about or any need or focus on generating money it's likely going to be very specifically focussed.