The problem is that the first-mover is going to have to be the identity providers: Facebook, Google, maybe GitHub or LinkedIn for specific markets. Users won't trust some piddly little startup with their authentication.

And without them, no developer would use it as their sole login (who'll use your service then?) and no users will set up accounts (who's going to do that for 0 services?)

Persona works by proofing you "are" that email. Your email provider needs to adopt this. If you run your own mail server you can do it yourself, if you don't, you can register with mozilla as a fallback.

For most people that means either Google (gmail) or some other party that will run email just for this purpose (for example, Facebook creates an email address for each profile already).

The percentage of the public able or willing to run an email server is tiny. That is not a solution for any market big enough for anyone to care about.

I expected that the amount of developers wishing for persona would hint a large shop to bootstrap the network effect.. alas.

The problem is that the existing identity providers don't stand to gain much if they switched, but stand to lose their control. Lose, lose for them, thus they didn't switch.