I never said they shouldn't improve aerobic conditioning, but it will not help if their AWC isn't there to begin with or isn't addressed. Unless you have a natural smashing AWC then you need to do some kind of specific training that involves deriving energy from anaerboic glycolysis and/or performing specific intervals that can work to decrease the rate of glycolysis per a specific time period at a constant power output then you're missing a piece of the puzzle. Yes, training your FTP will help to increase oxidative phosphorylation and delay the the rate of lactate oxidation by moving the upper bounds on a blood lactate curve upward, but that's only one piece of the puzzle because you still have to have either economy or efficiency from fast oxidative sources. Also, you can move one's OBLA upwards via different pathways on the power continuum depending on the nature of your events, however, training at or around one's Vo2 max can generate just as significant of a physiological response.
Moreover, it also depends how an individual derives energy in the first place and needs to for their event. Take this table below, for example:

How could you honestly know where on the continuum one needs to improve their energy systems without actually knowing which factor is their weakness? Sure, lots of people have shitty FTPs, but lots of triathletes have great FTPs and shit AWCs and probably would get dropped from a category 4 criterium if they weren't careful. In fact, last year a local rider managed to go from category 4 to 1 in about 3 months. As a former triathlete he had been taught FTP, FTP, FTP that's all you need for years, yet his coach actually had him do no FTP work after January and he was smoking fields from April until August. He spent the 3 seasons before that eeking out top 20 finishes and almost gave up on bike racing completely after having 3 different coaches put him on the FTP above all else type of plans. His new coach took him into a lab and blood tested him and found that he had such a steep blood lactate curve that his functional vo2 max was only 12w above his FTP. After spending a full 6 months of targeted L5 and above work he widened this gap by over 100w. So far this year he has done 3 total FTP workouts and recently received an invite to the Nature Valley Grand Prix by winning a pro qualifier event as an amateur. Quite simply, he has an amazing natural FTP, but doesn't respond to intervals over 8 minutes in length. He also does AWC work almost year round, even during the winter when others had told him to just do SST and FTP all day long.
I'm a great example of this case. Early in the season I tested with a 320 FTP, a 406 5 min, and a 430 2m. My fatigue profile tested 'average' and 'below average' for values from 5s out to 2min.
3 months later after loads of FTP work, my FTP moved up to 360. After 2 blocks of vo2 max my 5 min max increased to 420 and my 2m to 445. If you take that by percentages, my percent of FTP for 5m and 2m actually went down despite my FTP going up significantly. Performance wise I could drop many people on longer training rides or steady state climbs, but I wasn't finishing any better in races. Even with specific training, I still lacked the ability to recover from surges until I specifically trained my AWC and did stochastic vo2 max work. My 5s-1min. values stayed almost static. After 3 full blocks of AWC work, my repeatable 2min. bumped up to close to 500w and my 5min. up to 432.
Back to MTB training, I the MTB racers I know all do quite a bit of aerobic training on their road bikes, but in terms of specificity, they also do a lot of AWC and neuromuscular work. Microbursts with high cadence, microbursts with low cadence, L4 and L5 intervals with bursts, stochastic intervals based on terrain, cadence, and power, and even loads of LSD. All of them race on the road, but often will ride to the race, race, and ride home or they race on/off on the amateur elite/domestic pro circuit. Yes, all of them have high FTPs, but they also have absurdly high vo2 maxes and tested strong in both regards right out of the gate as juniors. But, I'm sure if FTP is all that mattered they could've just submitted power tests instead and development teams still would have brought them on....
I'd take a look at your local race series or ride series, talk to some riders, and take it from there. Once you can figure out exactly what the courses require, break it down a bit to plan your training. If there are lots of sharp, short hills then do some AWC or L6 work and FTP work with bursts/attack simulations. On your more aerobic days work in some technical work. Its a lot easier than it seems and if you can find someone with power files its even easier. Doing 2x20 at 95-105 RPM might help, but if that's all you do then you'll be SOL when you do an 800w 30s start and try to settle into FTP while hitting short, sharp rollers at 500w and 45 RPM.