Jens,
Why do you say net neutrality has anything to do with price? It's about best-effort delivery of packets without censorship or, for example, treating packets that say "do you want to join our radical fundamentalist agnostic cell," the same way as we treat packets that say, "do you want to subscribe to our newsletter."
In 1973, the packet switching X.25 systems which resemble today's internet more closely than the IMP point-to-point testing at the time had no provisions for packet inspection or "quality of service" adjustments. But if you didn't subscribe to a database that you might want to access (which may or may not cost money) then you had no access because if there were no login credentials then you could tell everyone how to use the database when it could only handle on the order of dozens of users at a time. What you want in saying that you think zero rating violates net neutrality is the MIT open Multics movement, which exists on the internet today in the form of free and ad-supported hosting services like Wikia. Net neutrality is about no preferred qualities of packet delivery service, because those are best handled by adaptive rate coding at the application layer, which is what the WMF causes the implementation of when they contract with cell carriers to allow access to Wikipedia content for no charge. The fact that Wikipedia is civilization's best summary of accumulated knowledge so far is the reason why carriers are willing to provide the transmission power to their users at no charge in areas where they still ordinarily compete on a per-bit fee. That is an economic application design choice that has nothing to do with packet delivery choices.
Similarly, in the 1860s the Hayes printing telegraph ticker tape had no restrictions on who could send a transmission or what it's content might be, and in cases of congestion, the operator noticing a collision first would back off, and the other would re-transmit in an egalitarian fashion, but the data you sent would obtain a response in proportion to the amount the recipient was being paid.
Wikipedia Zero is a great program and I hope something like Wikiversity Zero assessments will be how hundreds of millions of people learn new facts pertinent to their lives and helpful to them in ten years. With adaptive instruction coupled to Wikipedia Accuracy Review, I believe that such a system will support the transition from creating new articles to maintaining existing content. I hope both the WMF and the WEF support this effort, because if the WEF was paying for it, it would likely not influence the safe harbor provisions protecting the WMF from legal liability due to inaccuracies. I am sad when dictatorships use Wikipedia Zero for propaganda purposes, but I am not sure how much of a problem that is relative to the advantages.
Best regards, James Salsman