Mostly because the "world record" link they cite is not 2.4ghz and uses carefully-aligned directional antennas on both ends and "high sensitivity" receivers.  Putting that in a cubesat and broadcasting to consumer grade hardware, without an antenna or tracking, stretches credulity a bit.

The closest precedent I've found is the CanX-2 satellite, which was 3u and broadcast in the S band at 256 kbps, though likely to a directional receiver.

http://www.utias-sfl.net/nanosatellites/CanX2/system.html

One of the other issues they have mentioned is that the 802.11 spec requires a minimum speed of 1 mbit.  So there is still a gap between what is clearly possible and what they claim to want to do.