Vennemann doesn't really have anything else than some toponymy and base-20 counting system to support his hypothesis.
While yes – toponymy is indeed substrate – simply seeing a sequence as basic as "ar(a)n" in different places and claiming it must come from basque word "haran" (especially: note how his Val d'Aran, Arntal, never have that initial h- which is IMPORTANT in Basque – "aran" is a plum) is hardly an argument. The rest of the claims are mostly the same thing too.
But, yes, base-20 counting is a clear sprachbund thing. Same with the couple words he names that are shared with Celtic and Germanic – they're mostly related to iron age technologies, so likely wanderwords related to spread of tech (just like how industrial tech terms all over the world can be traced to various European languages).