Yeah, I'm paying at least twice that much here in Japan.

No wonder everybody can afford those ridiculous computer rigs with such low energy prices. :P

im in the cbd and pay like 22c for firt 1000kWh then like 24c for kWhs after

I'm in the outer suburbs, looks like it's 22c off-peak which isn't as bad, but still way off from the study.

And if you play on the gamepad you save another 400 watt for tv and surround sound system :)

PC's actually use bugger all power when idle. My Core i7 3930K CPU will range from 15 watts to 150+ watts of energy.
All my GPU's switch off when not in use with the primary GPU powering down to just 10-20 watts.

When playing a game though... All bets are off. :) GPU's will easily drive up to 600-800watts alone.
However, even if a CPU has a higher energy consumption rating than a slower CPU, it can actually end up saving you money in the long run as the CPU can "Hurry up and get to idle" far faster, saving more energy overall.

Energy is 30c/kWh here though so it's not cheap, but if you can afford a high-end PC, surely you can afford high-energy prices anyway.

However do keep in mind, the Xbox 360 and Playstation 3 on launch were certainly not as energy efficient as the later revisions, so expect power consumption to come down as the years go by and software updates are rolled out.

I can't afford a high-end PC. *sigh*

Actually the names of nodes have nothing to do with the size of anything, and in terms of rough tech equivalence, 20nm IS 22nm. Half nodes no longer exist. Also Intel 22nm is superior to TSMC and GF 20nm because it's non-planar.

And then the interesting thing is going to be that TSMC 16nm and GF 14nmXM will not actually be much smaller than 20nm, but cost a lot more.

Not to mention Samsung advertises it's "nodes" wierdly.
I.E. It advertised it's new "10nm class technology" which puts it anywhere from 10nm to 19nm. (That's for NAND.)

You are indeed wrong on the part that half-nodes no longer exist though.
Half nodes are those that are not defined by the ITRS or "International Technology Roadmap for Semiconductors""

For instance, 65nm, 45nm, 32nm, 22nm, 16nm and 11nm are classed as full nodes as specified by the ITRS.
Half nodes are 55nm, 40nm, 28nm, 20nm, 14nm and 10nm.

In general CPU's from the likes of Intel are always on a full node, sometimes the chipset or other chips may be on a half-node like in the case of AMD's 780 and 880 chipsets .
Traditionally SOC's from the likes of ARM, and in some cases... AMD APU's as well as GPU's are usually on what is called the "half node".
Intel historically used to keep it's chipsets on a node or two behind it's latest fabrication process in order to give the old fabs something to do as an idle fab is wasted money.

Right now the consoles are using 28nm or a "half node" eventually they will move to "20nm" in the next stepping, that probably likely won't happen untill 2015 however because of yield and power characteristics.
nVidia and AMD will likely sample it first and take the hit on their high-end GPU's untill the kinks are worked out.

And everyone knows Intel is the masters at playing the game of fabs. :)
They won't be beaten on that front any time soon, it's literally the only thing that's allowing Atom to go up against ARM.

I don't doubt that one bit. But people care about numbers. And 20nm is classed as a half node. :P

Besides, different fabs have different levels of transister density even at the same node, for instance TSMC at 32nm might pack less transisters per mm2 than Intel would at 32nm.
As a process matures, you can also pack more transisters into the same space.
There is actually allot more to packing transisters into a chip than the "nm" rating.
But to claim TSMC 20nm isn't on a smaller process than Intel's 22nm.. I would need to see some solid linkage, rather than take the word of what is essentially a stranger on the internet.

Not necessarily.
It depends on the fabricator's position.
The large cost is in the R&D (And building the plants!) you can bet those who are in the fab game does the R&D at every single step. I.E. 19nm, 18nm etc'.
That way they can observe such things like electromigration, leakage and mitigate it as much as possible.

I agree.
AMD won't have it's new GPU untill late 2014, by then the process would be new.
It would take a good 6-12+ months for it to mature and by that time, then it might happen. (In the case of TSMC however, never have your hopes up for yields to improve quickly.)