Every indication is that A10X is NOT TSMC 10nm. We will presumably know in a few days when Chipworks opens one up. But the Geekbench numbers suggest that it's the same process as A10. (Runs same frequency as A10, just w/ 8MiB L2 and 3 cores.)

The original source of "A10X=10nm" seems to be the (somewhat unreliable) Digitimes.
It gained plausibility given the delay in shipping the new iPads, and seemed reasonable --- if Apple were willing to dual-source the A9 to both TSMC and Samsung, maybe they were willing to perform a similar sort of "dual-target-process" trick for TSMC's 16nm and 10nm? It also gained plausibility in that the timelines lined up with when TSMC said 10nm would be ready.

But if you look at the news stories through Google, it was never widely agreed upon, or widely sourced. (Compare to the large-scale agreement by everyone that A11 recently entered manufacturing on TSMC 10nm...)

I think the most reasonable interpretation is that we all let our hopes excite us too much! And that the delay in iPad Pros from say October or so till now reflects not a delay to get the A10X on 10nm, but a delay for other reasons. (Perhaps sourcing the screens, which are definitely above average... Perhaps just a deliberate decision by Apple to try to shift what revenue they can from the pile-up quarter of Q4 each year when new iPhones come out into something more evenly spread? Perhaps an assumption that to get people excited about iPad sales again, they needed to launch the new models with the promise of the iPad-specific OS improvements that they demo'ed at WWDC but which weren't ready until recently?)

This all MAY imply trouble with TSMC 10nm? Or may be meaningless.
Certainly the Snapdragon 835 numbers suggest what I would consider only a disappointing process improvement over say the 821. This may reflect that Samsung 10nm is not that great, or just that rushing to be "First!!!" on a new process is kinda silly if the cost is that neither your design nor the process are really optimized.