We have to consider what benefit a prospective new payment processor can provide to our merchants right now as a tried and proven solution, not take a gamble on what potential benefit they might be able to provide later on.
As for PayPal, they have incentives to make it worth our while, and really most of their new services we're supporting use some combination of the same basic IPN and API specifications that we've already been using for Website Payments Standard and Pro, so it's actually been fairly simple for us to tweak our existing programming in the minor ways necessary to support these new services based on the same technology. Besides which, PayPal dominates the online payments market so completely that we can't reasonably afford to forego supporting their most popular payment solutions.
As to your final question, I can't speak to general usage stats Web-wide offhand, but we can provide a ballpark idea of relative numbers/%-share volume that our own system handles with each processor. Google Checkout is probably PayPal's closest competitor, which handles less than a tenth as many orders through our system as PayPal; Authorize.Net handles about the same as GC in the past year-to-date (but only half as much in the prior year), and each of the others we support handle about another order of magnitude smaller volume with us than that. Comparing the past year-to-date of volume against the year before: PayPal has done slightly larger numbers but about the same overall share in the YTD; GC's down slightly in numbers with a smaller share; Authorize.Net has gained ~50% in their numbers with us and gained slightly in their overall share of our volume in the YTD, so now they're about even with GC in both metrics.