As some may know, i joined RoboVM a few months ago. The team’s currently working hard on getting a 1.0 release out the door. We’d like to get feedback and thoughts from you folks on where you think we should take RoboVM. If you can find 2 minutes to spare, we’d be very happy if you could give this survey some attention. We can of course also discuss things here.
Thanks for any input.
Relatedly, I’m maybe even more interested in what RoboVM will achieve if released from the shackles of iOS. AOT compilation is awesome.
Cas 
I’ll probably use it in the future if I ever want to make my game engine for iOS…
Other platforms would definitely be cool. Especially would love to see a solid solution for Java on the web. As plugins are no longer cool, maybe something like RoboVM -> Emscripten/Asm.js. The web may also be the most lucrative in terms of the RoboVM business model.
You looked at TeaVM or Bck2Brwsr? I know they exist - I have no idea whether they are a “solid solution”. 
Is LibGDX GWT output not working for you?
Kev
As it stands, getting anything working in GWT requires jumping way too many hoops. GWT uses Java source code (as opposed to java bytecode) which removes a lot of flexibility when using pre-compiled third party libraries. From my limited testing CPU performance is also pretty crappy for anything mildly taxing.
Further there is a whole tool chain of C/C++ libraries and engines being ported using Emscripten but using them through GWT requires glue code which is another pain, a project like RoboVM could potentially allow one set of JNI code to be used for both native applications and web application which use those same libraries.
While the LibGDX folk have done really well to overcome some of these problems and made life easier for developers in the long run the GWT backend is still probably a pain to maintain and a such a tool could make it easier.