The creator of xwOBA, Jeff Simmonds, claims that LA stabilizes (in other words, eliminates randomness within the sample) after 70 events. Ian Happ, since June 1st, has 124 events. He is beyond the burden of proof of xwOBA by nearly an additional 50% and his xwOBA over that span is essentially, right in line with what he has done the last three seasons.
Speaking logically, it wouldn't really make sense for it to be the case you are suggesting and Ian Happ still have better than career barrel%, because barrel% takes ideal LA into account as well.
Between "it's just normal bad variance" and "maybe he has such a weird launch angle on this ground balls and fly balls that somehow his BABIP is really bad" the former is the far more likely and normal reason. Literally every year, across baseball, there are multiple hitters have two months of bad luck, just like hitters can have two months of good luck.
The more I dig into the data, the more it's almost assuredly that Ian Happ has been the recipient of poor variance for a while. If he continues to hit the ball the way he has, he will likely play in line with normal Ian Hap the remainder of the year.