His xBABIP is based on his line drive/flyball/groundball rates and if his BABIP exceeds his xBABIP, it's probably too high. Speed can certainly influence it as well, although Hernandez, while not slow, probably won't see a huge boost from his speed unless he starts hitting more groundballs.
For nearly every player in MLB, a high BABIP of .376 is usually an indicator a player is just having a great year they are unlikely to repeat. This year, only 3 players in MLB had a BABIP as high as Hernandez over a full season - Avisail Garcia, Jose Altuve and Tommy Pham. And for two of those guys, this year certainly looks like an aberration against the rest of their careers. A lot of other players have come close to his number, but again, many of them, such as Chris Taylor and Tim Beckham, are also having career years. Over the last 3 years, no player has a BABIP as high as Hernandez had last year. DJ LaMahieu comes closest, and he has an obvious advantage.
Hernandez has 116 career plate appearances in MLB, and those are spread out over 2 seasons. There is nothing he has done so far that should make anyone think they have an accurate picture of him going forward.
But if you are such a large proponent of these small sample sizes, at what point do we acknowledge his offensive fWAR of -2.6 is a negligible improvement of Pedroia's -2.7?