John Battelle noticed that Google has ... a lot of money. $10 billion and growing. Cash in the bank. Wow. They're experts at storing, searching, and otherwise wrangling massive amounts of data. They're investing heavily in wireless networks, data centers, and bandwidth.
I appreciate their acquisitions and efforts towards building and acquiring web-based productivity apps, but what the computing world needs more than Yet Another Office Suite is a cheap, persistent, and ubiquitous utility for storing and crunching data. Amazon is definitely on the right path with their Simple Storage Service and Elastic Compute Cloud services, but Google has the monster infrastructure required to make it truly pervasive.
What happens when metered computing and storage is accessible from anywhere? It removes the long-term commitment to a hardware investment, and becomes cheap enough to put in the hands of the little guys: A high school's computer club can build and learn about clusters of virtualized machines for a few bucks a day; digital artists and film makers can batch process and distribute their projects for pennies on the dollar; universities in developing nations can quite literally build ad-hoc supercomputers. And that's just what's possible now with Amazon's services.
So, I'd like to see Google step up to the plate with a little competition for Amazon. Utility computing is an emerging market, and Google has the mindshare to bust it open. They also have some of the brightest people in the business, one of the most capable computing infrastructures in existence ... and money to burn. How about opening those resources up to the public? I know I'd pay for it.