I have been managing my University moodle sites for the last 5 years. I have always done an upgrade each summer with the latest version - ie the May release, as an August upgrade. But I will have been using the master dev version right through development, including running on my own local dev environment and then on a development server that mirrors our live site. I also take part in the QA process, usually using my mirror dev site to run tests as well as the Moodle HQ QA site.
Installing 2.9 this summer (ie 2 versions behind) would mean that bug fixes are not supported even before the academic year starts and security fixes are dropped during the academic year
Bug fixes for general core bugs in 2.9.x will end 9 May 2016 (12 months).
Bug fixes for security issues in 2.9.x will end 14 November 2016 (18 months).
Likewise even 3.0 will run out during the academic year
Bug fixes for general core bugs in 3.0.x will end 14 November 2016 (12 months).
Bug fixes for security issues in 3.0.x will end 8 May 2017 (18 months).
So the options become LTS (currently 2.7) and dealing with the fact that we don't get some of the new features for upto 2 years) or a balance between early adoption and losing bug/security support.
2 versions behind is an option that many take, but its not entirely clear cut that its the best option - and 2 versions behind at installation, becomes 4 by the next upgrade, installing 2.9 this summer means 3.3 will be out before the next upgrade. You benefit stability and security at the front end, but potentially lose it with no bug/security support by the back end of the time frame (unless you also do mid-year upgrades, which I don't think many institutions are keen on)
So, yes early adopter, but with caution and full testing.