Heh, I noticed this news, too. It seems to me like there was never a really good implementation of a transputer, or at least there wasn't one that showed off its true potential to home computer users. Even on the industrial/research side of things they didn't seem to gain much traction - or, rather, not to the extent noticeable by average users/enthusiasts.
I am curious to learn how the old transputer model compares (conceptually) to our modern, multi-core, multi-processor desktop/laptop computers of today.