I believe "improved capabilities" != "new capabilities" to some extent. :) I also believe VP8 support would rather be available in Kepler with a lower process than the current Fermi chips that require a larger die. So the improved capabilities could be 4K (or larger) support and improved error resiliency, or other postprocessing features.
Originally Posted by deanjo
VP8 now would probably mean to use Google's code block as is. Assuming it was released earlier this year, that'd be rather quick for a GPU design. Rockchip already presented HW accelerated VP8 decoding though. On the other hand, new codecs were generally implemented in entry-/mobile chips first. Just wait for a couple of months then. :)