So far we've got confirmation from Nvidia that memory pooling isn't built into NVLink per say but the developers will have the option to implement it themselves.
To quote somebody from the GPU groups on other sites:
"Memory pooling is possible for GeForce RTX according Nvidia’s Director of Technical Marketing, Tom Peterson, during HotHardware on their 2.5 Geeks podcast:
Petersen explained that this would not be the case for GeForce RTX cards. The NVLink interface would allow such a use case, but developers would need to build their software around that function.
“While it's true this is a memory to memory link; I don't think of it as magically doubling the frame buffer. It's more nuanced than that today,” said Petersen.
According to Jules from OTOY NVLink memory pooling is going to be implemented in Octane 2018.1. I think ChaosGroup has hinted on doing that as well in some of their blog posts.
It is not a 100% but it does seem fairly clear still.
