I think NMC came very early before it has to and wasted so many years without proving its usefulness. First, are people seriously thinking of decentralized DNS now? I may be wrong but "not really".
In my opinion, the primary reason why we're not seeing a massive acceptance of NMC as decentralized DNS is the lack of usability.
Imagine all browsers supported NMC out of the box (e.g. including a light client which does DNS lookups) and you could register a name without hassle using a light client or even a web app. I think in this case it would be a serious competitor to centralized registrars - as a much cheaper way to provide the same service, and without having to deal with centralized services.
But we have not even a light client which can register names. That everyone should run a full node is a nice vision, but unfortunately, it's not realistic at this point.
And this is a chance for NMC investors - above all those that know a bit about coding, as they could simply port light clients, provide electrum server services, code browser extensions etc. It's also a risk, because another coin project can provide the same features - and maybe faster.
But I'm currently seeing some progress on these issues, and thus I've bought some NMC.
You are right that the most hyped projects at the moment are smart contract-driven "platforms" for "dApps", but a coin doing one thing and doing it well - and above all: really decentralized! - should fit at least into the top-50 coins. Most of the "infrastructure" platforms are also pretty centralized, what makes me wonder if some coins like EOS really have long-term viability - I think they will have a hard time once the "blockchain hype" has faded away.