I would multiply that number at least by 100 for the end of 2014
It is interesting to ponder how many ATMs will be deployed in the wild by the end of 2014. Lamassu has sold 100 now, and they appear constrained by manufacturing, as opposed to limited by sales. From my experience, I know that the next 500 machines are usually a lot easier than the first 100, so I expect they will soon be able to ramp up production quite effectively. Add in RoboCoin and some new players I've yet to learn about, and I think 1,000 machines by this time next year is possible.
If the ATMs average $1,000,000 of sales (the RoboCoin in Vancouver did this much in 29 days [albeit in the middle of a
bubble growth spurt]), then this means ~$1,000,000,000 of new demand. My rough estimates suggests a 5 - 10X multiplier for the effect on market cap. For this reason, I would imagine the bitcoin market cap to increase by > $5,000,000,000 (holding all other variables constant) should 1,000 ATMs be deployed across the (western) world.