On the surface, Google is the perfect candidate for it. Their infrastructure was made for this kind of stuff. From Bigtable to GFS, Google was made to handle web applications with little thought on the programmer side. And that’s why it’s so appealing to programmers. No longer do you have to worry about RAID drives, backups, load balancing, provisioning instances or servers, data storage, file serving. . . You just have Google take care of that for you. As you need more resources, the App Engine provisions them (within your limits which I assume will have paid levels in the future).

But could there be something a little sinister going on here. Most programmers hate worrying about those infrastructure details so Google is likely going to get a lot of interest from the Web 2.0 crowd. Of course, choosing Google’s App Engine makes it hard for anyone except Google to buy you. All of a sudden, your code is written for Google’s infrastructure and not a more generic Linux setup.

While migration is possible, think of it this way: if you’re on Google’s infrastructure and your product has an objective valuation of $3M, no one will buy you for that price because of the cost of moving to in-house infrastructure. So, maybe you get $600k? Of course, since you don’t have to move anything, you’re worth even more than that $2M to Google and Google can purchase you for the bargain discount price of $950k.

This seems like the perfect way for Google to be able to buy cheap startups and block their competitors from buying the Next Big Thing.