There are a lot of companies that are scraping Google quite successfully. Many of these are for 'rank checking' services that provide ranking data for certain keywords over time; these are heavily used by SEO and marketing agencies.

The two that jump to mind are Authority Labs and SEOmoz.

I guess: a shed load of proxies. :)

Amazon/ other clouds out there. Just auto provision your instances (lots of them), scrap, sleep, wake, scrap, sleep...