There are a lot of caveats to the stats.
First, look at the browser marketshare table.
Notice how Net Applications shows Chrome only at 16% while the others are close to 40%? Wonder why?
Net Applications attempts to measure what individual users are using, while Statcounter tries to measure how much of browsing is done though a particular browser. That is, if you browse 1000 webpages a day and use Chrome, and your girlfriend browses only 50 but uses Firefox, Statcounter counts you as 20 times more marketshare than your girlfriend! In this hypothetical scenario, Statcounter will give Firefox 5% of marketshare and Chrome 95%, even though half of people use Firefox.
This is like setting up a lookout on a road in a typical US city and concluding that the Ford Crown Victoria is the most popular car in America, based on the fact that you see it most since the taxicabs and police patrol cars drive back and forth all day(since you're not tracking the license plate) while the best selling cars like the Honda Civic/Accord, Toyota Camry/Corolla are not driven as much during a typical day. So, if you're writing a Chrome app, you're targeting only around 16% of people, not 40%.
Google bundles Chrome by default with Flash, Acrobat and Java updates which are installed on 98% of computers by paying a lot for it, Mozilla has nowhere near the huge ad revenues like Google does. Not to mention how Chrome is pushed on Google properties like Gmail and Youtube. I keep seeing a "Slow browser? Upgrade to Chrome." message on Gmail sometimes while using the latest version of Opera.
The biggest challenge Chrome apps will have is that they won't run on the hottest devices right now, the iPad and iPhones. Apple prohibits any sort of native API or dynamic code in iOS apps so Chrome apps will be confined to HTML5 and WebGL.