What? Do you seriously believe that PBMining is NOT a ponzi scheme? You cannot be serious.
At one point a few months ago, the amount they were charging didn't even cover the power needed for two years after buying the cheapest possible hardware. However is the S3's out those same calculations probably don't work. Still they wouldn't have funds to upgrade the hardware.
I think the business model of PBmining depends on Moore's Law holding true. You're buying a mining contract, not a share of specific equipment. As time goes on, 1 GH becomes cheaper to buy and requires less electricity to operate.
I think the business model is viable without having to resort to ponzi methods. It's my understanding that the mining equipment is outsourced in a few locations around the globe. The locations likely have very cheap electricity. They even have a competitor that charges slightly less.
Could the business model fail? Sure it can. It makes certain assumptions about the future, and those assumptions could be wrong. I would only invest a small amount of my portfolio in any one cloud mining company (and mining in general).
Create a spreadsheet, plug in the numbers, and decide if cloud mining is worth the risk. I think the more important question is if you think cloud mining is profitable at all. With the current cost of contracts and the difficulty rising an average of 16% each time, I'm waiting to see what happens.
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. 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.
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.