Is it? What factor should one apply to those number to get the real cost?
You can't. You have throw away that number and compute a valid number. It costs about $2k to buy 1 ghps, which gives you a 1.3x10^-8 chance at every 25 btc block reward (one every 9 minutes). That difficulty increases 20% at each adjustment, every 12.8 days. Now calculate.
"now calculate" - hehe, yeah, right.
as a miner for 3 years now, I would say that my opinion of the "cost" of mining will closely correlate with the bitcoin price moving forward.
there will be significant lag, as there is a lot of sunk money in bitcoin mining, but my greatest hope would be the graph below:
credit to TERA for this nearly famous chart.
such a scenario would dampen mining in a profound way. we really had a great push when asics were first realized, and then implemented, and now they have been executed.
Now we are reaching the boundaries of what is currently possible. The physical boundary of heat is IMO the biggest barrier at this time.
Until the next "quantum leap" in computing, mining bitcoin is going to get quite boring. (it already pretty much is)
altcoins keep it fun, kinda, and I can see the real innovation in trying to find new super-secure, energy-efficient, wipes-your-ass-for-you types of encryption.
the only problem is that with 5 years of accounting records behind it, bitcoin has a large incentive to stick around...
... sorry, I started out trying to pick a fight and now I don't even remember what I was mad about. isn't the internet great?
Granny was young once and got a boyfriend 
. OTOH, Chinese officials are very inconsistent - mobile websites still OK. Truly, it will be kinder if they would just openly and honestly shut everything down, not suddenly closing access to online wallets or off shore exchanges. Guess these are cultural differences, for some reason they do not want to appear to ban btc, even if they are effectively try to do it 
