I think Google considers their infrastructure and expertise with web-scale data processing to be a durable source of competitive advantage. You need to be substantially as good as Google to do web-scale crawling and analysis to build a general purpose search engine. Google makes billions upon billions of dollars from having that available. Are they going to make billions upon billions of dollars with a huge competitive moat and 30% gross margins by publishing their competitive advantage as OSS and getting into hosting? Probably not.
Which is not to say Google isn't a friend of OSS and transparency -- indeed, they are the world's most reliable fans of OSSing the competitive advantages of companies which are not Google.
> "Are they going to make billions upon billions of dollars with a huge competitive moat and 30% gross margins by publishing their competitive advantage as OSS and getting into hosting? Probably not."
But are they going to make billions upon billions of dollars when their competitors have instant access to a service that levels the playing field?
[edit] Okay, not quite level, but even with Amazon collecting their margin, it dramatically closes the gap in highly scaled computing, to the point where if I were Google I would not rely on cost advantage as a long-term strategy.
Competing on price is rarely a game you want to get into, it tends to turn into a race to the bottom more often than not.
Which is not to say Google isn't a friend of OSS and transparency -- indeed, they are the world's most reliable fans of OSSing the competitive advantages of companies which are not Google.