You're spot on with the convergence of lighting and shading models, and that it presents new opportunities for asset sharing.
Without getting into any specific titles though, I'd like to point a few more, pretty important facts:
Both game, cinematics, and movie VFX asset pipelines can have practically the same first steps: after the 2D concepts are finalized, multi-million polygon high res assets are created in Zbrush, and all that detail will get baked into texture maps. The only difference is that games will have a low-res poly mesh and normal maps, whereas a movie asset relies on subdivision and displacement, and can usually afford higher texture density.
(for this reason, sometimes there's a secondary layer of displacements painted in a software called Mari (funnily enough that's basically "Mary" in hungarian))
So, the same Zbrush sculpt can be used as a start for both an ingame and a blockbuster movie asset, and then the pipelines will only diverge at the technical steps. Yes the game asset uses a fracture of the texture space and poly count of a movie asset, but they are created with the same tools now.
(I'd also need to add that sets aren't sculpted that often, and normal/displacement detail can be generated in the form of general maps (like a brick wall texture) instead of a full Zbrush sculpt of a huge building. But this is once again becoming very similar in movie and game workflows.)
Also, Digic is still just about 110 people and that includes everyone from backoffice through production management and animation to the actual asset creation team. Blur and Axis aren't much larger than that either. Schedules are also very short, we rarely get more than 2-3 months of time to build stuff for a trailer. So the amount of content we can produce is actually not that much, and a very very small percentage of what needs to get into a nextgen game. Oh and we also fake whatever we can when the camera allows it and use matte paintings not only for backgrounds but even for characters when it's possible.
On the other hand, internal game development teams are sometimes as big as 300 to 500 people, with entire studios built on outsourcing in Eastern Europe and Asia that even the largest publishers sometimes rely on. And all these armies of artists have years to build content.
So the actual lesson is that it's more and more possible to share certain assets
both ways, that it makes a LOT of sense to leverage not only CG assets in the game, but also ingame content elsewhere. For example, Ubisoft now has an internal team called Passenger that's focused on exporting ingame content from the engine to be used in transmedia productions. They've been formed for the Assassin's Creed Lineage web series and have been involved in several projects since then. I wouldn't be surprised if some ingame content would find its way into the AC movie, the big one with Fassbender.
http://gdcvault.com/play/1015579/Passenger-Story-of-a-Convergent
It also means that nowadays it's a must to to keep an eye on game development and look at all the GDC and Siggraph publications from game studios. If anything, the hardware restrictions are making devs even more inventive and there's a lot to learn from them.
