A recent study looked at how medical marijuana legalization has influenced alcohol consumption. The study is solid, with falsification and robustness check and consideration of covariants and confounders. According to the researchers, medical marijuana legalization reduces consumption by 15%. This is the first real evidence I have seen that marijuana is a threat to wine. At this point I would characterize the situation as 'a preponderance of evidence indicating that marijuana reduces ALCOHOL consumption.'
The logical conclusion is that this would reduce wine consumption specifically. On the other hand, my much less robust research points to evidence that (a) as marijuana use has increased over time in the US, so has wine consumption; and (b) that states where marijuana use is high also have high rates of wine consumption.
Let me be clear: the study cited before easily trounces my bit of number-crunching. It is a much larger step toward a conclusive finding. On the other hand, it does not look at wine specifically, but alcohol, in general. It has also not yet been replicated or peer-reviewed in a serious manner yet, as far as I know. Smaller producers should also note that it is based on scan data, which focuses on the largest brands, at the largest outlets, and accounts for only 40% or so of the market.
If I had to guess based on the evidence I have seen at this point, I would assume that (a) legalization will reduce total alcohol consumption by 15% or less; (b) legalization will reduce wine consumption by significantly less than 15%; and (c) that small-production, premium wineries will see even less of a hit. Whether or not the hit to wine results in a detectable reduction in sales for all wine and/or the premium segment remains to be seen.