I don't have a good answer but I wouldn't rule out luck and statistical variance. Countries tend to go through Golden Generations followed by lean patches. Maybe we just didn't fully realize that those 1982-1989 birth year cohorts didn't only represent a leap forward in player development (although that was clearly part of it) but also riding the high side of variance, with negative regression inevitable afterward.
Variation is probably part of it.
It's also fair to say that 1982-83 in particular was a HUGE bumper crop (Donovan, Dempsey, Beasley, Onyewu, Convey, Beckerman) that papered over underproduction in other years and that also caused the USSF to draw some erroneous conclusions (e.g. "Bradenton is good!").
1982: Donovan, Beasley, Onyewu, Beckerman, Gomez, Goodson
1983: Dempsey, Convey, Clark, Wondolowski
1984: Guzan, E. Johnson, Parkhurst
1985: Cameron, Kljestan, Holden, Feilhaber, Findley, Evans
1986: Davies, Spector, Edu, Zusi, Orozco, Castillo
1987: Bradley, Bedoya, Besler, Torres, [F. Johnson], Ream
1988: Gonzalez, Lichaj
1989: Altidore, Adu, Villafaña
There's a decline in among players born in the 1980s, but I'll chalk this up to variation. The best players from the late 80s (Bradley, Altidore) aren't as good as the best players from the early 80s (Donovan, Dempsey), but it's hard to draw massive conclusions from that.
But then the early 1990s are just so much worse and while I'm cognizant of
@coremiller's warning about overfitting a theory, check out the best players from this generation:
1990: Nagbe, Hamid, Corona, Shea, Hedges, [Chandler], [Diskerud], [Dwyer], [Morales], [Johannsson]
1991: Zardes, Garza, Rowe
1992: Wood, Lletget, Agudelo
1993: Yedlin, [Brooks], Trapp, Zimmerman, Parker, [Saief]
1994: Morris, Bono (this year is sooooooo bad for some reason, there aren't even many decent MLS pros)
If you made a best XI from all the players I've listed above based on their primes, the only US-developed player from the 1990-94 group who would make it is Yedlin. Granted, 1982-89 is eight years and 1990-1994 is five years, but the ratio is still bad.
I'm sure there's some luck involved, but there has to be other things going on as well. The shittiness of the Bradenton model & its coaches, changes with the DA, endemic complacency among to players...I don't know.
1995 has depth but not much high end talent. It will get a boost if Miazga and Steffen can emerge as NT first-stringers. 1996 is bad. 1997 and 1998 are good. 1999 probably won't be epic but should be okay, maybe better if a few players hit the jackpot on their potential. 2000 looks very solid. So we are on the upswing.
Given the population and resources we have, there's really no reason why a solid year like 1985 shouldn't be the bare minimum. Crank out two EPL starters (RIP Holden's career), and some players who had their moments like Sacha and Benny. I believe that you can't control when the truly elite players will come around, but you can control the baseline and you can create the conditions so that elite players reach their potential when they do appear on the scene. There's no reason we should have years anywhere near 1994, when there just aren't many players who are even decent at soccer.
Ultimately we need the soccer powers that be in this country (and I'm not just talking about USSF bureaucrats) to acknowledge that even 1982 type years are just scratching the surface given how many kids play soccer. Can you imagine how good the US could be if we had even 1/6 of the productivity of Iceland or Uruguay?