Better investment (in my view) is buying bitcoin and holding it for the next 6 months at least.

If mining would be likely to yield bigger returns than the price of the machines, most asic providers wouldnt sell their golden egg laying chicken, instead they would mine themselves. Or if they are smart, do both and screw their customers twice.

What many people dont grasp is that the variable production cost of these mining chips is on the order of $10-$20 a piece for the bigger 28nm ones. They are selling them now for $1000-$2000 or more. Until demand will dry up (due to the difficulty increase), and then those asic companies  just drop their prices and will keep dropping them, which will cause D to keep increasing exponentially for the foreseeable future, rendering those early asics useless and unprofitable. This will only stop once mining profitability crosses over with variable production cost, and my best estimate is that wont happen with 28nm chips before we reach somewhere around 1 exahash:
https://bitcointalk.org/index.php?topic=286126.msg3143855#msg3143855

Now you may get lucky by betting on the right horse and getting your hardware soon and while most of the competitors stumble, but overall miners will bleed lots of money as these asic companies are just auctioning off their risk to the highest bidders (/idiots). Dont get involved for profit, but by all means, if you are curious, put an old GPU or one of those USB mining sticks at work. Just dont expect a ROI, you stand far better chances winning on satoshi dice.