I think the long-term valuation largely depends on how good the content is. Not to be too negative, but the content right now is not very good. I see the successful posters right now as highly paid beta testers more than legit authors.

My thesis is that steem could get into a continuously cycle of "better
content-> higher price -> better content -> higher price".

A reference point for what good content costs is maybe the COGS of the New York Times. At this level we would have a reward pool of $650 Million and a market cap of ~$6.5 Billion to support such a pool.

Add a speculative premium to that and you get a steem price of around $50.

My prediction is when the trending articles are the same quality as NYT journalism you'll have $50 steem.

It could go several times higher than that, but ultimately everything depends on content!