If there's anybody that actually thinks that conspiratorial, consider two mutually exclusive options, one of which must be true:
1) Companies collecting massive amounts of data will only use the data in a lawful and ethical fashion, even in cases where they perceive there to be a potential for significant gain in misuse.
2) Companies collecting massive amounts of data will, sooner or later, use that data in an unlawful or unethical fashion to exploit others in a way that stands to potentially significantly boost their power/influence/wealth.
I'm going a far step above beyond and suggesting the exploitation will come, not from leaks, but directly from the data harvesting companies. The reason is simple. Again, the two scenarios above are mutually exclusive and one must be true. So what would you place the probability weighting as? I think most of everybody would agree that the probability of #2 is very near 100%. It's also possible that it is literally 100%, in that it has already happened but we don't know about it. Truly effective blackmail would obviously not be headline news.