The problem with this is the word "potentially". For those of you who don't know, we count the number of planets round stars and discern their atmosphere by measuring the changes in light coming from the star. They count the number of planets by looking out for periodic dips in the stars light, but it's not very accurate. They have no way of knowing whether two dips of similar but differing size are two planets or one on an irregular orbit, or if a dip is a planet or a particularly dense part of the systems asteroid belt. Similarly they have no way of knowing if it's one large planet or two approaching synchrony in their orbits. They figure out the elemental content of a planet by looking at the changes in light received from the planet. Essentially, all atoms have their own photonic barcode they stamp any light touching them with. When passed through a prism this light refracts into its different colours, revealing the barcode. [url=http://www.astronomyknowhow.com/pics-res/hydrogen-spectra.jpg]Here is hydrogens barcode[/url], to make it easier to understand. Again, the problem with this method is we can't tell the difference between a large earth like planet, and a small gas giant with a Venusian moon, or two close by planets each with half the requirements for life. So whilst in [i]theory[/i] there could well be 8 billion habitable planets out there, chances are that a large amount of them will turn out to be wildly inaccurate false positives. Uneducated guess- 1 billion (generously).