Sure, I just wanted to put the actual numbers out there because 20% or so actually seems pretty fair for keeping the lights on, taking out the trash, etc, while

Sure, I just wanted to put the actual numbers out there because 20% or so actually seems pretty fair for keeping the lights on, taking out the trash, etc, while 70+% is....a lot. On top of this, a lot of the....infrastructure is still fee-for-service. The university may fund the initial purchase of an MRI scanner (or whatever) out of overhead, but individual labs still pay $400+ an hour to use it. I don't know if the NIH or NSF takes overhead into account, but many of the most successful grantees also have some of the highest overhead so probably not. On the flip side, I've heard someone claim that their department chair strongly encouraged applying for NIH grants, because the NIH-negotatied overhead rate is much higher than other funders (also, allegedly more prestigious, but you'd think $500,000 spends about the same, regardless of where it comes from).