But, as one of my clients would attest, even if you're a paying customer, you may still be left hanging in the wind. I hope they don't start spreading the YouTube feedback treatment to all their services.
If they spent 15% of that ($2.1 bil) on support, with say a mix of 2/3 low pay ($100k/yr cost) and 1/3 high pay ($150k/yr cost) people... they could hire 18,000 additional support staff and still show a great profit.
That works out to 30 mil support hours assuming 1680 hrs/yr (35 hrs * 48 weeks) actually spent on support. They could probably hire twice that many people by hiring some people in less expensive areas.
To sum up, Google puts making a more money ahead of better support. It's their right to do so, but it not "unrealistic" for them to provide much better support.
I'm not disputing that they could improve their support, just the price tag and scale that you have placed on it.
If someone figures out how to automate support better, they're going to get paid.
At the outset of Google's growth, their products and APIs seemed timeless; with the passage of a little time, that's changed.