* The biggest incentive for a gadget developer might be the large community Google (and thus, the Google Personalized Homepage) has. Recent statistics show some high page view numbers for certain gadgets. Thus for some developers, it may be enough to just reach Google with their gadgets.

* While a universal widget framework may be a nice bonus, any abstraction layer, like this framework too, comes at a price (the more layers you add, the more potential for problems, and the more layers you need to understand & learn - deciding which new technologies to learn and which to ignore is a constant conflict for every developer). Now developers simply need to decide if they receive enough benefits to justify paying this price.
As an example: even if the Unversal Widget API supports every feature the Google gadgets API offers today (I doubt this), then what if Google introduces a new gadgets feature tomorrow? Well, then Google gadget developers will already be able to use it, while UWA developers will have to wait two weeks (or so) to see this feature being available in their framework. (Again, this kind of lag may be justifiable depending on the bonus you get from using UWA.)

* Already, you can create plain (D)HTML code and then include them as gadgets on other sites, without changing your source (the Google Personalized Homepage allows you to embed URLs).