I rent a server for $500 a month, with $100 in electricity. (easy numbers, not an exact quote of course.)
My costs every month are going to be $600 total just to mine bitcoins. This is a flat rate, always going to pay it whether I mine bitcoins, or run a Quake 3 server.
I can make $1000 a month mining bitcoins with this server. That's a $400 profit, after taking out electricity and rent.
Now, if I can get someone to rent this server from me for 1 month, and pay me $600 a month for it, and only expects $700 in bitcoins, then I cover my cost of the server, the electricity, and I still keep $300 in my pocket. I'm not going to pay them $1000 worth of bitcoins to rent the server, I'm going to hold some back.
My guess is that if they are charging 0.25 for 3 days on their server, then over those 3 days, they are making 1.0 bitcoins, and only giving back 0.35, netting a profit of 0.65. All with no risk to themselves.
Now, I could retype all of that substituting Bitcoins in the small denominations like .02, .05, .10 and what-not, but it's 6:35AM in the morning, I've had a half a cup of coffee, and crusty eyes, so I don't have the mental power to do the calculations.
Now, I'm also going to try selling Premium memberships... That's just cash in my pocket above and beyond what I'm making. Who knows, maybe I'll throw a couple of ads on the front page to make a little extra. The ads will help, because no sponsor wants to work with you upfront unless you already have an established user base. Just like any website, you have to get the users first, before you can get the money.
That's how I see it.
Pretty much this is how it would come to pass. You buy a house, but rent out 80% of it to pay the mortgage and get a little extra (if you can convince people it's worth it). Over time the house is paid for, you have equity, and you could reinvest.
These people could, over time, build more rigs as they build equity. Those that 'invest' get a 9% return, and when no-one is invested, they mine coins for themselves. They're dealing with all the hardware issues, any downtime, and will have a pool in case of severe downtime to pay off those who have invested at a small loss to the website.
It's a pretty smart idea. Obviously only invest in as much as you can afford to lose, just in case.