Is it really that surprising that Google finds crawling pages stored on their own CDN to be faster than pages stored... anywhere else? They don't even need to leave the Google network.