Right now 10,000 GH/s should be paying around 0.187767 BTC / day. I see at PBMining it is .577198, okay it is Tuesday night in the States, but .577188 / .187767 = 3.073. So PBMining has paid out 3 days worth and no deductions for power!!! Okay the claim is the power was paid for up front.
When you buy the contract, it is stated that all fees are paid for up front. This is not a guess or speculation, but a statement made by the company.
That may be possible but doesn't seem likely. I know at Hashnest.com you can get 1 GH/s for .00135 BTC. PBMining is charging .0029 BTC, a differenct of .00155 BTC ~ $.74 USD. Let's assume they only pay $.03 / KWH which is super cheap by any countries standards. That means they can buy 24.67 KWH. Right now the 1 GH/s burns about 1W if you have really good hardware. (It has to be cooled too, etc) that means there is only enough power for 2.81 years.
I thought that really good hardware was closer to 0.5W/GH. Even Butterfly Labs has equipment that reaches 0.7W/GH.
Let's try a mathematical example:
I buy a 1 TH contract for 2.9BTC. At today's exchange rate of about 475 USD / BTC, that is 1,378USD.
Equipment is probably 0.60USD / GH, or $778. That leaves 600 USD for power and profit.
At 0.5W/GH * 1,000, that's 500 watts. In a day, that's 12kWh or 0.36USD. In the first year, that's $132. That leaves 468USD for years 2-5.
In year 2, they will have new equipment. Let's assume it is 0.10 USD/GH and power is now 0.25W/GH.
Equipment cost 0.10USD/GH * 1000 = 100USD
At 0.25W/GH * 1000, that's 250 watts. 6kWh/day or 0.18USD. In a year, that's $66. Now there is 302USD left over.
In year 3, they will have new equipment. Let's assume it is 0.05USD/GH and power is now 0.1W/GH
Equipment cost is 50USD, electricity is 100 watts or 2.4kWh/day or 0.07USD/day or 27USD a year. Now there is 225USD left over
Repeat for year 4 and 5, and you still have money left over.
I hear Iceland has cheap electricity, and cooling is nearly free.
This is actually much better than the last time I checked, but still far short of workable. My assumptions don't leave any room for anything else like their cut.
The problem is simple - The payouts are impossibly high with no variation like you get with real mining and there isn't enough left over to cover expenses. Moore's law doesn't matter, but there isn't even enough money to buy power, let alone replace the hardware.
It just depends on your assumptions and calculations.