Github data is not "gargantuan". They have 9 million users. They put limit to repository size to 1GB. My estimate would be that most repo don't even get to 10% of this. Plus this is mostly source code and text files that compresses very nicely. Given this information, I would estimate that Github data can fit in under ~5000 2TB hard drives. Cost of hosting data would be around $5 million per year. I bet their employee cost is much larger than this (my estimate is $50M).