I still have no idea why we're learning this from the Windows press rather than an honest and informed Microsoft spokesman, but Paul Thurrott has an article on his site today that digs into some of the potential reasons for the power management debacle:
https://www.thurrott.com/mobile/microsoft-surface/64867/trying-to-explain-the-surface-sleep-problems
His article does implicate the Skylake chipset, and flags the point that Microsoft is the first to try to make this power management method work.
The bad thing is that the article pretends that all is fixed, when it quite clearly is not - at least for a great many of us. On the other hand, my machine woke correctly after being put to sleep with the power button today - time to celebrate! (Except
that two desktop apps were crashed on wake: Firefox and Sublime Text - I'm not sure whether those are Win32 or x64 apps - I'll have to look...)
There's one more datapoint for at least some of the remaining problems being related to the keyboard cover - The cover was not present when I put the SP4 to sleep, and it was removed before I woke it successfully. (BTW, both the SP3 and the SP4 have always
been a bit flaky with the keyboard cover folded back out of the way (at one point, booting the SP3 with the keyboard cover folded back failed or did something bad - I don't recall what. I expect there are subtle bugs in the keyboard position sensor/state
code that have been hiding out for a couple of years now.)
It's pretty sad when we get excited about something like wake/resume just working as it should, though. Further, I'm quite disappointed that MS isn't owning up to this and communicating with us themselves. It sounds as though they really are trying to
offer a best-in-class experience, it's just not working yet. (Boy, is it not working, yet...)