I guess you'd need to crawl the web and look for those headers. Or maybe you could look at IP addresses? It would certainly be difficult to do with any kind of accuracy but you could probably get some decent estimates if your sample size was large enough.
Rendering web pages is actually one of the worst use cases for Amazon from a bang-for-the-buck perspective, especially when you factor in bandwidth.
When people are building 30,000 core compute clusters [1] on EC2 - presumably with zero publicly available web servers, I'd be very interested in any methodology that provides reasonable estimates of revenue based on public web servers.
http://arstechnica.com/business/news/2011/09/30000-core-clus...
That's likely only if you use Amazon's Linux distribution (based on CentOS). My EC2 instances say "Server: Apache/2.2.17 (Ubuntu)".
(You probably have to have showdead on to see me because I'm hellbanned. I've emailled pg to try and get this fixed, and had no response. If you happen to see this, please check my comment history to realise that I'm not at all a troll, and consider upvoting me on the offchance it will get my account back into positive karma land. Thanks!)
Somewhat out-of-date (Aug, 2011) but a useful estimate.