correct - the mining process is scraping over your plot files.
The amount of storage devoted for plot files is equal to your chances of finding a block.
It also defines the amount of CPU power you need to scrape, because you need to read 1/4096 of your plotted volume. In most cases this is not important and can easily run on a single core, in the background, not hurting other processes running.
The actual "plotting" involves a CPU-intensive process and may take a day for a Terabyte. The i7-3930K I used for plotting yielded 2 TB/day. Lots of people buy a couple 4TB disks and don't want to wait 2 weeks to plot them, hence the development of a gnu-plotter.
Its the same problem. I am buying 1000 TB and it will take weeks to plot. Having only CPU i will not think of getting 1000 TB. GPU only make people scale up and everything end up the same but people have to spend more money.
1000 tb are you serious?
That is only 250 4TB drives. Ok 275 4 TB Drives after microdick finishes messing with them.
lol "only", you need multiple rig for that, and you can't dump the amount produced by that, not worth it
right now even 100tb are not worth it