The comment you replied to answered that: Because of the hypothetical situation that Google would want to pay $5bn for it. You'd buy it to shop it around to the big players.

To a certain extent, the valuation of some companies can be pumped on the basis of to what extent the big players are likely to be competing for it, and not always on the basis of possible revenue, but strategy too. E.g. denying a competitor a company that is seen as having some percent chance of being a threat if paired with the competitor, or buying a company that is seen as having some percent chance of helping you become a threat.

Consider that even if Snapchat on its own was/is worth 0, if SnapChat has even a tiny chance of helping Google propel G+ into the position of a viable contender to Facebook, Snapchat is suddenly worth a huge amount to both Facebook and Google. It might even be worth a lot to Google even if Facebook only thinks that the combination would be a threat, if it'd make Facebook tie up lots of resources trying to "counter the Snapchat threat".

(in that vein, I think that one of the ingenious part of the massive Gmail quota was that while it got quite a bit of PR attention that mostly swayed geeks, it terrified large parts of Yahoo - I was at Yahoo when Gmail "hit" - and it tied up hundreds of man-years worth of product manager, engineering managers, developers, marketing people and others radiating out from the Mail team, all trying to figure out how to respond; I can only imagine the same must have been the case in Hotmail and others; as a feature Google could've offered a much smaller increase over the competition and still gotten most of the PR, but that huge bump, which was largely symbolic, triggered massive resource hogging responses from competitors)

And a lot of that kind of thinking boils down to guesswork, which can work both for and against them in terms of valuation.