Basically, Musk is alleging is that they claimed this was a common practice when it was, in fact, extremely rare.
In his tweet about this he said that out of 5.5 **billion ** ad impressions that day, less than 50 were objectionable according to Media Matter's criteria. In other words, there was a 1 in 100 million chance that a normal user would randomly see something like this.
For comparison, the following things have about a 1 in a million chance of happening (i.e. are 100 times more likely):
- flipping a coin 20 times, getting tails every single time
- winning the PowerBall lottery if you buy six tickets a week for a year
- a devastating earthquake occurring in Seattle within the next 5 hours
I just read the MM piece and it doesn't appear to make any specific claims about how frequently this might have happened, it merely says "We recently found ads for Apple, Bravo, Oracle, Xfinity, and IBM next to posts that tout Hitler and his Nazi Party on X." and that "X has been placing ads for Apple, Bravo, IBM, Oracle, and Xfinity next to pro-Nazi content." which does indeed appear to be factual since it makes no claims about frequency, so I guess we'll see if the court is convinced that it was defamatory. It certainly seems to be the truth, but not the whole truth.
If it turns out they really DID have to create 100 million page views in order to find a single questionable ad placement, and they failed to mention that, you could make the case that they were intentionally trying to hurt his business.