Correct--where I said "IP" above, I should really have said "patents/patent rights". Copyright, to my understanding, becomes more of an issue when publishing, as some (most?) journals are greedy in that regard, but I haven't had to cross that bridge yet.
What you're describing sounds pretty similar to the way it works here, but the gross majority of full-time graduate students in engineering are employed by their universities (as "research assistants" or similar) and thus effectively waive their patent rights.
(FWIW, I think it's great for schools to share in patents which originate within them, but I would like to see students and staff be more well-informed on the subject, and commercialisation efforts which span beyond a single industry--I know of very few schools which have accomplished that.)
I was pretty sure this policy of students/academics own their own copyright but share in patents was the standard in United States higher education.