In the case of captioning the videos, the captioning costs several times more than putting up the videos in the first place. Probably on the order of 100x more. A lot of these lecture videos are just someone with a camera recording the lecture live, someone who would be there, watching the lecture, regardless. They do some trivial editing / splicing, and upload to YouTube. This is essentially free, other than having the professor/school sign off on them doing it.
Once the ADA gets involved, you now have to caption the video while watching, or after the fact, something you likely would not have done otherwise. You probably also have to submit it to some authority in the university now (don't want to get sued!), who will have to rewatch the video and verify the captions.
In the old system, it took you less than 5 minutes of inconvenience per lecture to upload them. In the new system, it takes you 2 x len(lecture) => 120 minutes or so, per lecture. This is a huge tax in relative terms.