Integrating graphics card into their chips is what Intel seems to want to do as well.
So it would appear AMD was ahead of the game.
I guess, at the time, the thinking was that GPUs were a core threat to the CPU business itself, since a lot of the buzz was around using the GPU to do CPU activities.
The PS4 is an example of that. It's an APU with GDDR5 acting as both system and video memory.
To be honest for a while I thought that was where everything but servers would be going. You can tune your OS and applications around the slower GDDR5 and it seemed a no brainer: instead of 8GB of system memory topping out and then 4GB of VRAM doing jack shit you could have 12GB that could be dynamically assigned to whatever you wanted!
Alas, that dream was never meant to be.
Edit: I posted about this a while a back and apparently Arch Linux has a hacky way to use your VRAM as system RAM. Good stuff.
Zen in many ways is still moving towards that model, The concept of an IO die that all cores hook into, along with all of PCI-E lanes, no need for complicated cross core linking. It would be relatively "easy" to introduce specialized accelerator cores or beefy GPU chips into this mix on a separate chip, and not massively complicate the CPU design, or massively bloat its cost as a specialist die.
So it would appear AMD was ahead of the game.
I guess, at the time, the thinking was that GPUs were a core threat to the CPU business itself, since a lot of the buzz was around using the GPU to do CPU activities.