Dunno what the fuss is about.
I updated to v9 right after it came out, going from v8 on all my rigs and there have been ZERO issues. A little more power use and spikes up about 40-50watts on the 5-6 GPU rigs, but nothing that can not be handled.
I run my own BIoS on XFX based hardware and it's been stable since v9 came out, without any crashes etc..
I always say, check your hardware first as more than likely it's a power supply, cable or motherboard issue.
Just because the v9 taxes your whole system more and IT can not handle it, doesn't mean it's bad.
My r9-3xx/2xx GPU's added about 12-18% in hash rate, and the RX based GPU's about 4-5%.
Sadly with the price of ZEC, I might have to switch back to ETH soon.

ZEC is a second life for my old 280x cards. ETH is for my new RX

And about the fuss - electronics components do age, manufacturers use barely passing QA components, etc....
In my last batch of RX cards it was a miracle to find a card capable of doing 2100 memory on eth. Tested v9 on them and i had to reduce memory clock to default 2000 otherwise rig rebooted.
My old 280x cards were running 1625 mem clock on ETH mining. They could mine ZEC with those settings with initial Claymore releases but with version 8 I had to set memory to default 1500 on all of them. With v9 I had to reduce 1 card to 1400 MHz for stable mining.
Blaming "power spikes" is easy. More important is having discussion about the real causes and how to overcome them.