This is such a good move! It gives Google the nimbleness, hunger, and guerilla mentality of a start-up in new areas it wants to explore through these small mostly-autonomous teams, while simultaneously allowing it to defend the already captured beachheads (search, gmail etc.) - all funded by the deep, deep, Google pockets.
In any innovation-oriented org, curious engineers and inventors need to be able to play and push the boundaries, but even large organizations with strong financial backs are so defensive when it comes to innovation, so afraid to fail, or waste resources on experimenting. Google has always been okay with this "waste". If you go back before year 2k and try pitching to a goliath sw company to let 20% of dev time be spent on employees’ projects of choice you'd get assaulted by the CFO. Google was okay with this "waste", because they knew if you let the right players roll the dice, every now and then you’d hit jackpot. And they did! Many of their most successful products came out of the 20% project.
Organizations today have split the vision and execution aspects of building something. The vision comes from management and the execution from engineers – this is straight from the defensive playbook - ‘engineers can execute with minimum risk, and managers are close to the customer therefore know what will sell for sure’. This kind of thinking will work when you want to improve marginally (like Henry Ford said something along the lines of 'If I asked my customers what they wanted they’d say a faster horse'), or if you are the market leader, but it will never cause disruption or let you make headway in uncharted territory. It is very important to know when to play offence and when to play defense.