I think it's the other way around. The architecture was being limited by their fabs ability to yeild large chips and in the absence of any CPU perf pressure fro

I think it's the other way around. The architecture was being limited by their fabs ability to yeild large chips and in the absence of any CPU perf pressure from AMD the natural push would lean more towards increacong graphics performance in order to push more pixels. As in I think Intel probably had the same yeild issues as everyone else ~10-32 nm but only Intel had the high margins small chip volume to make it profitable to ramp until Apple and TSMC happened.