When Google started, Beowulf clusters were all the rage. So it wouldn't be a big conceptual leap at all. But productionizing it is serious work.
Money you get by being a pair of connected Stanford grads living in the most tech connected place in the world, with Stanford alumni and venture capital connected people all around you.
The CS is part is medium-hard. The productionizing part is a hiring problem.
Getting the capital is a whole other aspect.
I knew lots of smart people in 1996 in the first .com wave. None of us made any money :-) I should have moved to San Francisco in 97 like I was originally planning. Oops. (And no, I could not have done what Jeff Dean & Sanjay did but I would have loved to have tried ;-) )