This has to do with calling a single command and blasting out an app for IOS once you have already written an app for Android.
If you note carefully Google could have done it the other way around, by implementing IOS shims on top of Android.
This would have been easier (as Google controls Android) and because Objective C code is low level code that is easier to port to high level Java code.
And it would have given an incentive to those who have already written IOS apps to port them to Android.
So why not do it that way? Because then everybody would develop for IOS first and Android second and few people would use the Android specific features. By going from Android to IOS, Android is the primary system, but Google still lover the cost of development of Android apps by having the IOS market articficially subsidise it.
It is the classing Jole post (http://www.joelonsoftware.com/articles/StrategyLetterV.html) with the twist that Google is lower the price of the complements to its complements (ie. smart phones are complements to Google ads and apps are complements to smart phones).
I haven't built ICS apps, though, so perhaps it's getting better.