It appears he is suggesting that Intel's competitive position with 14nm will be worse, presumably because the ARM forces will have 20nm by then.
But why would someone think that ARM will get a larger jump in the 28nm to 20nm transition than Intel would get in the 22nm to 14nm transition? If anything, I can accept the notion that it will be a wash - but even TSMC's own data suggests that their 16nm FinFET process gets a full "traditional" node of power/performance/density improvement - but 20nm gets the full density improvement and half the power/performance, and 20->16 gets the half of the power/performance improvement.
Intel is aiming for a full node improvement - so potentially, we can see Intel increasing their lead with 14nm.