But it's for the long-term health of their company.
There are too many business majors out there running things who don't seem to grasp the simple difference between long-term and short-term gains so they do things that piss off their customers for a short-term profit and end up paying for it later. And frankly, why wouldn't they? The whole reward/incentive scheme in business is completely set up to encourage this. You get a bonus for an idea that results in a profit, and the focus is always entirely on the next quarter--so why would you come up with an idea that's going to increase profits in 10 years when you may be working somewhere else. Hell, for that matter, why wouldn't you come up with an idea that is less profitable over 20 years but more profitable for the next 2 years? You get paid more if you do.
Google--I think, rightly--feels that it is strongly to their disadvantage to be a company that just does one thing, which is ultimately what they are. Companies that just do one thing tend to do fantastically well--until they die horrible, terrible deaths. Look at Blockbuster. Look at the print industry. Look at Circuit City and now Best Buy.
For Google, these current billions are pocket change. Sixteen billion dollars over the next ten year is just 1.6 billion a year. They can find that in their couch cushions--for now. But if they wait until their luck starts to change to figure out how to diversify, they may not be able to spare the cash to do it and they might end up going the way of buggy whip makers and everyone else who just did one thing. All of these things are gambles, but they're calculated gambles. It's ok if they have a 99% chance of going nowhere, if the payout for the ones that do go somewhere is way more than 100 times what you put into it--at least that's ok if you're a company like Google with obscene amounts of cash-on-hand and a desperate need to diversify.