Planning 6 months out in the bitcoin world is nearly the equivalent of planning 6 years ahead in the outside world if you ask me. Ignoring that, friedcat working with Allied Control on the cooling system shows long term thought. It will not be long before ASIC's hit the wall of the current state of chip development, and miners compete not on hardware, but on electricity and cooling costs.
Iceland and Washington state are going to be huge in the next 5 years.
It doesn't mean BitFountain just be a "R&D > Deploy > Enjoy Good Times > 'Mining Famine' > R&D" and repeat. Less than a couple years ago, people thought ASICs weren't feasible and the thought of developing something that was more efficient than an AMD GPU was crazy. ASICs still has quite a ways before it reaches current CPU sizes (~22nm). DeathandTaxes had a post back, talking about BTC ASIC vs. CPU. While everyone is clamoring over Gen 3, here's something to consider:
Using
BitcoinClock as an estimate, the next having will be on 2016-08-25 (very likely sooner). Based on that timeline, we're looking at maybe 1 or 2 ASIC generations before it's down to 12.5 BTC a block? Preferably, it'd be nice to be ready for that.