These things are in development for many years before they hit silicon, I don't think his time at intel has been long enough for that to be true. More likely, IMO, is that the same bureaucracy that lead to intel falling behind drives people like Keller away.
If scaling pipelines is hard, scaling pipeline pipelines is harder.
He was probably there specifically to help Intel regain that edge. But he may not have received enough support to implement his vision, or simply he had a personal reason to leave.
If scaling pipelines is hard, scaling pipeline pipelines is harder.