The best networking I did in grad school was attending startup-related events. That got me my first paid gig, which led to greater street cred as a developer. Attending events held by the career center was a close second, because those led to interviews. Everything else was fluff.
I know that my PhD completely changes the nature of jobs I can hope to get - not really because of what I learned directly at my PhD, but for everything else.
Sure the majority amount of my/our time is spend sitting and working in the lab, but going to a few conferences per year, doing an internship most summers, and having weekly meetings with other researchers adds up to establishing a lot of contacts in the research community that I wouldn't have without graduate school.
Most PhD students -- at least the ones I know in the states and overseas -- spend most of their time in school trapped in a lab or office. A few make time to "network" with people outside the department, but most have to be forced to socialize with a broader audience than can be found at the weekly beer hour.
The far more common pattern for a PhD student is to spend years busting his/her ass in near-isolation to accomplish research goals and graduate, only to find that it's hard to get a job without real-world contacts. So you're absolutely correct in the sense that PhD students should be networking, but it just doesn't seem to happen very much in practice.