Here reverse DNS is being used to describe the machine having this address, not the services provided by it.
It is important that there isn't a 1:1 correspondence between websites and servers. The website "google.com" isn't handled by a single server, it is hosted on hundreds of systems, each of which has its own unique name to allow the admins to identify it – and at the same time, those machines host more websites than just the single "google.com" so it wouldn't make sense for their reverse DNS to point to just that single domain.
So instead of trying to put potentially inaccurate and rapidly changing information in rDNS, the operator just uses it to directly describe what sort of device the address has been assigned to. The hostname
zrh11s03-in-f4 was chosen by Google's infrastructure team to describe where the specific machine is ("zrh" is likely Zurich, the rest probably indicate racks or ports or interfaces or something similar).
The overall domain 1e100.net is a Google-owned domain name. It is common for network operators to have several domain names, with the ones used for public-facing sites being kept separate from those used for internal infrastructure. They're managed by different teams with different purposes and rules, etc.
Often the infrastructure domains are named after something which alludes to the company's name without actually mentioning any trademarks (for example, the network's AS number is a very common choice and you'll see a lot of
asXXXXX.net in traceroutes).
The number 1×10100, written as
1e100 in most programming languages, is called a googol.