I can think of two languages that were ahead of their time and so they may have been seen as being released early, since developers weren't ready for them.
The first is Smalltalk. It was one of the first OO languages, but there didn't seem to be any need for developers to change their paradigm, so it languished for over a decade.
The other was Erlang. It was developed at the same time as Java, but it required a paradigm shift, and since Eriksson wasn't ready to actively promote it, it's release was probably too early, so it is relatively recent in becoming more popular, as other FP languages become more mainstream.
There are some ideas, such as Cuil, http://en.wikipedia.org/wiki/Cuil, that should have been better tested and they should not have set expectations so high. If they had just come out, and had people try it, then, if it was better than Google, which they could have iterated before their was such a bad experience with it, then it would have spread by word-of-mouth.
Oftentimes it seems that new ideas will fail because of unmet expectations or the idea is too new and people don't truly see the use of it.
An example of that is the Wardenclyffe tower, http://en.wikipedia.org/wiki/Wardenclyffe_Tower, by Nikolai Tesla. If he had finished it we would have had pagers in the 1920's, but it was too far ahead of it's time.
I think there is less risk from releasing too early than too late. Generally, if an idea is ready then there will be several competing products in development at the same time, and if you wait too long, perfecting it, before releasing, then you may have lost your chance at a large segment of the market.
I think if you follow the customer development approach then, as long as expectations weren't set too high, and you can react very quickly to problems and reasonable feature requests, then too early won't be a problem.