That’s a big investment and a huge deployment… it should be enough reason to do some thorough hands-on testing before buying 100 devices But really, those devices don’t look like something you’d wanna run 30 clients on. I could be wrong, and I would recommend trying to look for experience with the new RPi3 because it has a broadcom WIFI chip that supports AP mode. Perhaps people have experience with it.
You can read more about it here: https://www.raspberrypi.org/forums/viewtopic.php?t=138730
If you find a dongle with the same chip, it probably supports AP mode. But you also need a linux distribution with the right drivers to make it count.
So basically, if you buy these dongles and expect to make them work over a wider variety of devices and linux distributions (or Windows!?) then it’s a lot of uncertainty.
I’m not sure if you acknowledge the difference between AP mode and AD-Hoc mode, but it’s significant. I don’t think you can reliably connect 30 devices in an Ad-hoc network, nor can I recommend that you hand over unconventional and sort of “hacked” environments to schools and local system maintainers.
Regarding access points, you can also buy cheap APs. For instance, I have hundreds of dlink APs from a former ISP. I mean, just look for them, and you’ll find them. They are much easier to configure!
It’s over the top to spend $10,000 on 100 APs anyways, if you would go for such a low-spec device instead… you would be very well off spending ~$20 on real APs from a certified refurbisher. Preferably, but not mandatory, with a router function.