This one is highly problematic without some kind of a Indo-European POD... which is entirely plausible. Not enough people on this forum seem to be keen on totally rewriting history. Lurking, I've seen a couple of attempts, but they all seem to die.
But yeah, Lusitanian may not have been a Celtic language - it's not very well documented, so we can't really tell. Was the Indo-European /p/ retained in earlier Celtic dialects? Our definition of Celtic today rests on the languages that we have the most data on, which lost that /p/, though it is entirely possible that it was something like /f/ in Romance languages early on, that is, a highly unreliable sound change on which to begin classifying linguistic affiliations within the family, since /f/ turns into /h/ and disappears in a number of Romance languages in different branches. Maybe the loss of /p/ was something that had become predominant in the family by the time of Roman arrival, but Lusitanian could still have been said to be closer to Celtiberian (which retained /kw/ in place of /p/, a la Goidelic but almost certainly independently) than to Gallic and Brythonic. If it was a Celtic language, then you just need to butterfly the Celtic invasion of Iberia... assuming Iberian was related to Basque. If it wasn't Celtic, and is in fact representative of a different group of dialects that invaded the area even earlier (perhaps related to some of the languages in Sicily, or maybe Venetic?), then you're going to need to butterfly Indo-European migrations... which is not ASB at all, I don't think, it just changes everything.