It's probably a decent idea to start "small" (if a billion dollars can be called that) and evaluate how well it works.

Apple's got enough money to experiment with multiple approaches and pour lots into the ones that work down the line.