This doesn't come at all as a surprise to me. Google Labs was well known internally for being a graveyard for projects that, ironically, was really hard to launch stuff in. Anything that Google launches anywhere automatically gets a huge amount of traffic, and, because it has the Google brand on it, it's expected to meet a fairly high quality, stability, and scalability bar. In practice, this meant was never really possible to use Labs for throwing something up quickly to see if it stuck. This, combined with the maintenance overhead of keeping the various projects running (which were often grafted onto other larger codebases), makes shutting it down a pretty obvious decision.