There are plenty of roadblocks, such as the mask manufacturing and inspection, stochastics and EPE (from shadowing). The assumed doses don't match the resists being tested. Even 40 nm features are shifted from their designated place on the mask.
There are not enough tools to do anything meaningful. In Q4 2016 and Q1 2017, 35 immersion tools were sold, this is more than the booklog and field EUV tools combined. And this adds to the already hundreds of immersion tools in the field. And we know EUV tools are slower. They can't be used on enough layers to make an impact.