Right. But it's not like a Threadripper Pro 7995WX with all its 96 cores has four times the multi-core performance of a 24-core M2 Ultra. Apple seems more interested in eventually trumping the performance with fewer cores than in offering products (to a very small niche) that offer many cores.
You're mostly right - though multicore workloads can vary *wildly*, with the single most important factor being how independently subtasks can run. The more independent, the closer to perfect scaling you're going to get, and the more the Threadripper/EPYC/Xeon will pull ahead of the Ultra. On a simplistic level, that's the key distinction between GB5 and GB6 multicore scoring.
The other key issue here is that you only have so many watts to distribute among all your cores. Intel and AMD have gone big here, with their biggest chips, drawing hundreds of watts. Apple has declined to play that game so far.
Apple is unlikely to "trump" their performance at the high-multi-core end any time soon. They're not interested in that market... at least, not enough to try something there yet. Perhaps with the rumored 4xMax-type configuration, if they ever release it. It's easy to imagine a unit with a single Max and three additional CPU/GPU units with anywhere from 16 to 32 P cores each, coming to 108 P cores (32*3 + 12) or more. But do they think there's a market for that? There isn't right now, but perhaps they will have the patience to develop it.
I think this is actually looking more likely than it was a couple years ago- the need for more data center processors may drive Apple to build their own, and if they do go to that trouble (as they have recently been rumored to), why not share those dev costs with a product they can sell, not just consume?