I don't really think, except in the case that you are legally required to store the raw data for decades, that there is any reason to store petabytes of data to recheck your conclusion. I'm not aware of any retention policy generally required of scientists to do that- nor do I think the value of being able to recheck a conclusion by investigating this kind of raw data is really required.
If it was required, you'd be willing to pay the (large, but economically justified) cost of archiving the data.
If it was required, you'd be willing to pay the (large, but economically justified) cost of archiving the data.