It hasn't bottomed out yet either:
Something to note as a matter of caution, though. Andy Revkin has taken a very careful approach on this, and reminds us that one paper found that modelling indicates that (suprisingly) intermittent positive trends in Arctic ice are still possible on 2 -20 timescales until the middle of the 21st century. This must be remembered; as any temporary recovery in the next decade will be claimed by the climate change denialists as proof positive that AGW predictions were wrong. (Just as they love to falsely claim that the 2011 Australian floods were inconsistent with climate change predictions.)
However, I would have to guess that this "reversal in trend" possibility is looking a bit improbable given the dramatic loss of volume in sea ice.
Modelling sea ice behaviour does seem to be a particularly complicated thing, though.