I tried searching the spec itself (http://hybridmemorycube.org/files/SiteDownloads/HMC_Specific...) but it doesn't seem to contain any specifications about the latency.

Lots of talk about latency-minimization though, but it seems this is basically a packet-oriented interface (with CRC on packets, retries and stuff) so I guess latency will be larger than with today's DDR interfaces.

Perhaps computer systems will have both DDR memory and HMC, letting the OS and/or applications decide how to distribute access for maximum performance.