You note that I always talk about MapQuest, but you haven't paid attention to what I say when I do. I've been saying for 2-3 years, long before the MapPoint API was introduced, that Microsoft was eventually going to eat MapQuest's lunch, because they were just thinking of themselves as an application, and not making themselves into a platform. It was after hearing one of these speeches, in which I argued that a platform strategy beats an application strategy every time, that Jeff Bezos made the commitment to turn Amazon into a platform. Unfortunately, the folks at AOL/MapQuest were far less receptive.
It's not just having an API, though. It's also having an architecture of participation. Amazon's customers, and EBay's, help to build the application by contributing their own data. Neither MapQuest, maps.yahoo.com (vicinity), nor MapPoint have made this leap. And it's such a powerful and obvious one for mapping/location data. I'm convinced that the player who first does this is going to be the one who wins in the marketplace. I'd bet on Microsoft, but it's not impossible that someone else will get there first.
In short, the API is an important step, but not the end game. In addition, it's important that there be a low entry point to the API. MapPoint's API, last I saw, was too expensive. (I don't actually know the pricing, but when the FAQ says, "sign up for an evaluation account and a salesperson will call you", you know that it's not cheap!) They ought to offer a "free for non-commercial use" license, or one that is free below a certain number of requests (like the Google API), rather than their current "free 45 day trial" model, if they want a lot of experimentation.
Net net, I agree that MapPoint is incredibly important, and the likely winner in this space. But there's still room for something disruptive to take over the market, or for one of the other players to expand their offerings. Until MapPoint starts getting buzz among the geowankers, I won't say they've got a lock on this market.