I'd argue that the Elizabeth Line was delivered in an Agile way, at least as far as is possible for a project of that size. "TfL Rail" ran the line from Liverp

I'd argue that the Elizabeth Line was delivered in an Agile way, at least as far as is possible for a project of that size. "TfL Rail" ran the line from Liverpool Street through Stratford out to the East, and I believe parts of the line out from West London to Reading. The first iteration of this was with old rolling stock, and then they even started running the new Elizabeth Line trains on the routes before the Elizabeth Line itself opened. This meant that by the actual opening they had proven out the trains, a bunch of stations, and a large portion of the line (by distance, not complexity), as well as trained many staff and operated it as an a separate group that ended up becoming the Elizabeth Line. Now they did still have to open 6-8(?) new stations, change the signalling, change the timetables, all on day 1 of opening, but there are still other ways in which they did an incremental approach. Trains didn't run end to end on day 1, they would run shorter, simplified routes, they decoupled Bond Street as an emergency exit from Bond Street as a station, etc. I'm not sure what more they could have done. Personally I felt the incremental "agile" approach worked here, kept them on target, and got value early on (I used TfL Rail for years!).