I would add a router to the list (suggesting the 1010 as it's most versatile when you have to stick to one out of budgetary reasons) as it can be used for edge processing like chamfers/rounding/patterns, flush trim for edge banding, slots/rebates, making recessed grips, copying from templates, mortices, with LR 32 for shelf pins/slides/hinge holes, put into a CMS, ... You'll most likely sink some money over time on accessories (like WA-OF, UP-OF, SF-OS, LR 32, ...) and bits but it's amazing what can be done with it.
Should you get a MFT with flip-rail: take the time to set it up from scratch (including adjusting the stop blocks for the rail holders in the tracks) prior first real use (make any test cuts with sacrifice sheet so you'll only cut into the real top after all is set and square). Easiest is to use bench dogs to align the fence and the rail, as the hole pattern in the top is square. The SYS-MFT-FX-Set is, while looking pricy, a nice companion to the MFT - but as you should get the MFT-SP (they're very handy to fix workpieces on the top for sanding or routing) and some more clamps anyway the cost isn't that huge, given the other parts you get with it.
Joining stuff: Domino is a unique tool and a great timesaver. Should that be out of budget but you have a router you could make yourself templates (for a copy ring) to precisely manufacture mortices for normal dowels or dominos. With creative clamping you could also create and position mortices for custom (wide) tenons using the LR32 system.
Sanding: Rotex - 150 if you want to process big areas, 90 if you need to go into corners. Have both and you'll most likely never need another sander/polisher (unless you do daily use or mass production, then a lighter/specialised sander might be nicer to use).
Rails: Always holey ones (to be ready should you ever get the LR32), skip festool rail connectors and get the makita ones instead.
Costs: When you handle your festools carefully their resell value will be great, sometimes it can even make sense to get a specialized tool for a one-off project and sell it afterwards.
Finally: always spend a bit of time and scrap material on a new tool till you're confident that it'll do what you want it to do.