What % of those demographics have bought Steem (i.e. have put any of "their often very low savings into a funky new cryptoasset in order to be able to use a blockchain)? What are the demographics of people who have bought Steem with non-Steem generated assets (so excluding SBD)? That would be the relevant stat.
To date, Steem has subsidized the creation of new accounts to enable new small accounts use of the blockchain, which is partially why the demographics skew to locales where that subsidy is more meaningful. The stats you shared say a lot about the viability of Steem and what has been accomplished here, but next to nothing about EOS, or DPOS in general.
Why? Metcalfe's Law doesn't care whether you bought into the network financially, only that you're part of the network.
EOS can have people contribute to the network via work, just like Steem does. It has a similar budgeting system. And merely being present on the network adds some value in its own. (Increasing the value of stake to others including the original holders).
In answer to "Why?": Because that would not specifically address Vitalik's point.
I do not believe that infinite access subsidies for the poor is part of EOS's design -- Dan conceded Vitalik's point on this issue, and countered that Ethereum is even worse for the poor.
I don't think he did concede that point. He went on specifically to say that services pay, not users.
Users including the poor will use EOS just like they use Steem where they will have greater barriers to entry on Ethereum and similar models.
Empirically Steem is being used massively in the developing world, including by people who are not wealthy at all. So contrary to Vitalik's claim that they need to purchase stake to use a blockchain built on this model, they don't. Services provide the stake (perhaps through delegation), and users provide value in other ways.
Fascinating observations.