Yes, I have experience with this kind of design. Mine was a 6-layer, 0.5 mm thickness, HDI with SBU. Total dimension 18 x 36 mm, it integrated 5 BGAs with minimum pitch 0.40 mm.
The game changer, apart from having ViP (Via-in-Pad) is the SBU (Sequential Buildup): choose the sequence wisely in order to obtain maximum flexibility but also avoid exaggerated costs. For example, on a 6-layer, you can have blind vias 1-2 and 5-6, buried vias 2-5, plus the usual 1-6. This sequencing is not that much more expensive than the traditional 1-6, but it allows you to do pretty much anything you want on the board, e.g. placing two large BGAs face-to-face on opposite sides of the board.
If you are not going for SBU, filled and capped vias allow you some neat tricks, not to mention the freedom you can take with the soldermask: bye bye solder wicking!
On the thermal side of things I haven't noticed a huge improvement, but my design was not that challenging on this aspect.