Well, we'd basically want to target them all!
Given that people change their phones more often than their PCs, one would expect the installed mobile browser base to 'fresher' than the PC one.
In other words, over the medium term (say 2 years) the rendering issue should hopefully fade away.
It's just a question of timing - when is it going to be OK to go 100% HTML5? (assuming, of course, the app is not particularly complex)
Ahh, I had assumed you were talking about a brand new application/service. If, as you mentioned in a reply to another comment, you're just exposing an existing web service to mobile phones then simply providing a mobile alternative to that service gets you a large percentage of your likely user base for a minimal effort.
If you're asking when 100% of the phones in your possible market will have browsers on them that are HTML5 aware/capable, I couldn't begin to guess. I don't think that will happen soon, but it's certainly happening at a faster rate than I ever thought it would be. This is the kind of question that can have a paralyzing effect however. If you want to provide a mobile gateway to your existing service you should go for it and you should implement to target the phones you expect the majority of your users to be using.
It was pointed out earlier, but it bears repeating, plan for what's out there right now. HTML5 has a lot of momentum in the browser space, but that doesn't mean it will be adopted in a uniform manner. Limit your scope if you have to, something is better than nothing, but right now the state of mobile browsers is in flux and I wouldn't recommend depending on 3rd parties to create a desirable environment for you.
I agree it's all about trade-off. Cleaner, faster, cheaper development vs risk of alienating some percentage of users with older mobile browsers. I was trying to get a feel for what people are currently thinking - but there doesn't seem to be an overriding consensus yet.
As regards focusing on what the stack looks like now rather than making an educated guess as to the not-too-distant future, here I have to disagree. Our business is all about well-reasoned gambles, and it seems pretty clear HTML5 will win the day for all but the most complex / intensive apps, so it really only boils down to timing.
It's just a question of timing - when is it going to be OK to go 100% HTML5? (assuming, of course, the app is not particularly complex)