Daystrom Institute

I always took the view that while the Ambassador class was a good, proven ship design which pushed starship innovation further for Starfleet (first ship with phaser arrays) it was costly and time-consuming to produce compared to its older brother in the Excelsior class.

Starfleet has a lot of time and experience building the Excelsior class, and likely has the production process all but perfected and can crank them out quickly and efficiently, unlike the Ambassador.

Then at some point, Starfleet started to mess around with a new era of ship that eventually cumulates in the Nebula/Galaxy class ships. During this process, however, the design RnD come up with some smaller but useful ships that are just as effective if not more so than the Ambassador in many different roles. These ships are the New Orleans, Niagara, Springfield and Challenger class ships.

It's inevitable that Starfleet is going to shift a part of its difficult Ambassador production over to these new ships that are getting the job done with less time, manpower and resources, not the easier to produce Excelsior. As a result, the Ambassador is slowly sidelined.

I describe it as the Federation taking a "peace dividend" at the end of the Klingon Cold War that disrupted new starship development for a generation or longer. Excelsior-class was the best platform to result from cold-wartime development and thus became the workhorses that lasted through the transition to the more "peacetime developed" heavy explorers like Ambassador and Galaxy-classes.

The real-world parallel are the US Navy Arleigh Burke-class destroyers. They were the peak design at the end of the Cold War, a peace dividend was taken which slowed shipbuilding at the time, their overly ambitious successor classes were ultimately unsuccessful, and it was decided to cancel the newer Zumwalt-class after only 3 ships and instead keep building more third-generation Arleigh Burkes.

It's entirely possible that there was also a technological leap made not long after the Ambassador was in service, much like what apparently happened with the Galaxy class. It might not have been that the Excelsior production continued at the expense of the Ambassador, but that there is a gap in ships where Starfleet didn't want to build entire new spaceframes that would be unable to take full advantage of new developments. Refitting the massive number of Excelsiors and Mirandas that already existed might have been the stopgap solution until the next generation of ships (Galaxy, Nebula, Akira, etc) was available.

The Ambassador class likely had a lot of teething issues that took years to fully correct and made it easy for Starfleet to sideline in favor of building brand new Excelsiors that have registries as high as 40xxx while Ambassadors only have registries in the 20xxx.

As for the original NCC 2000 Excelsior, background info for PIC S3 has Hikaru Sulu's ship decommissioned in 2320 and ending up as a museum ship at some point afterwords.

There's an USS Excelsior mentioned in TNG S7 "Interface" in 2370 that has a newer 5 digit-registry and confirms it isn't the original TOS movie-era Excelsior and presumably it's a newer-built Excelsior class ship commissioned in the 2320s.

https://memory-alpha.fandom.com/wiki/USS_Excelsior_(NCC-21445)

I assume the original NCC-2000 Excelsior has a lot of internal differences compared to later Excelsior class ships since it was the prototype for the class and those differences made it difficult to upgrade and refit and was the main why it was retired after only 35 years of active service.

A lot of starfleet technology by the 24th century is modular relative to the actual frame of the vessel.

We've seen ships eject and recover their own warp cores a few times in the series and shields emitters are able to be upgraded and become more efficient and powerful over time. The same can be said for phaser technology.

Nacelle placement and hull design have some impact on the maximum warp a vessel might achieve and for how long, but generally speaking this is at speeds that a workhorse vessel doesnt need to achieve and for quick bursts structural integrity fields should hold. Improvements in hull plating and armor composition also give us minor but important gains in taking hits when shields go down but this is less essential during peacetime.

In addition that the class was ahead of it's time. As a result of being designed to chase transwarp speeds the excelsior class' hull design was better able to adapt to 24th century warp technologies and wasnt dated as quickly as it's contemporaries.

So it was a good sized vessel, reliable, had an innovative hull design, and was easily able to be upgraded to have a faster warp core, more powerful phasers, and better shields and structural integrity fields. As an exploration and science vessel it was ideal, and if it needed to defend itself or aid another vessel it was still more than capable of holding it's own and then some.

It wasnt until the borg and the dominion war that the federation had pressures again to start pushing out more sweeping upgrades to their fleet and the result of that was a more battleship like combat ready fleet.

It seems like during the dominion war a lot of ships were taken off mothball status and retrofitted to join the front lines and I'd wager a good chunk of them were left in service after the upgrades and even found to be worthy of future retrofits. Like we see the Titan during Piccard is still a constitution class, but it is a very modern vessel.