Actual librarian checking in here.
1. I suspect the need to bring along an extra 99 books is a compromise between the inefficiency of 100-book chunks and the cost of purchasing and registering an individual box for every book in the library.
2. Changing from barcodes to RFID is one of those funny situations where smaller libraries with less overall funding are actually at an advantage. Remember that you can't just switch systems midstream and continue from there; you also have to retrofit each of the hundreds of thousands or even millions of items already in the system. That runs into big money fast, less for the RFID chips than for the additional labor costs.
3. Space is a huge problem for academic libraries. After a couple of decades you hit capacity, but your collection keeps growing forever. You're established on a campus where space is always at a premium, and you probably don't have the political clout to win a new lot when the engineering and business schools are also competing for it; even if you could claim the space, good luck in raising several million dollars for a new facility. Remove old, unused stuff to make room for the new stuff? We do some of that, but a shift big enough for people to notice will raise a shitstorm you can hardly imagine. When you factor in the fact that library's now see their mission as being as much about providing "space" as providing information, the problem grows even more rapidly.
I agree with you that the pricetag seems high for the benefit offered, but I haven't seen their actual financials so am willing to give them the benefit of the doubt. (I still personally hate robotic systems, for the record. Browsing is good, and just not replicable online.)