It reminds me on a company i once worked at, where the managers bonus payments were connected to delivery dates. Which more than once resulted in "roll software out now, add quality later" deliveries. This meant happy managers who got their bonus payment at the price of unhappy customers.
Yeah I had a project manager that quite literally said to my face: "I don't care if it's broken, roll it out, once it's live it's a support issue, not ours."
As for the news, i'd guess they realised they screwed the pooch, and the band-aids over the September patches need properly fixing. Pushing Fleet Carriers out would be too risky, and the realism (from the developers?) overcame management's reluctance to veer from the project plan, and after much shouting and heated meetings, someone found the balls to change direction.
Sadly i'm also guessing they haven't got more resources (from other teams / experienced ED devs working on 2020) so the only solution is to delay things, and hope team B can fix things and get Fleet Carriers out with the new timetable.
If things had gone smoothly in Sept, then one would hope Fleet Carriers would have been on track, although a 6 month delay is massive: if they were on track, and the extra time is purely bug fixes (with extra time to setup beta test environments and support structures etc) then the codebase must be properly buggered.
It's sad the more positive conclusion is Fleet Carriers weren't on time, the code isn't so bad but some months are need to fix Arx properly, and then a bit more for carriers.
The other possibility i can think of is that the 2020 update is whatever new features, but a branch of the core game's code, and it's been worked on for a year plus just refactoring, redesign and tidying the codebase.
So take the core of the 2020 code which relates to live Elite Dangerous, all new and tidy, remove/disable new features and pull in current features (which might not be in the 2020 branch). A faff sure, but you end up much tidier code, done by the A team, while also essentially testing the 2020 code early and effort is less-redundant as you don't have working on completely obsolete code.
This is in the realms of conspiracy theory though
