My own take is that the ef-s line up is pretty well covered, between the EF-s and EF lenses, especially given the likely budgets and aspirations of the aps-c market. Canons kit zooms between the 18-55 and 55-250 combo are pretty good, compact, light, cheap, the uwa could be faster, but how much R&D for how many units?
Canon have upgraded their ef short primes, which make great standard lenses on aps-c.
I think a wider cheap pancake, in ef mount to keep volume up and costs down.
Yep, Canon has the standard kit (18-55), standard tele (55-250), medium zoom (18-135, 15-85), faster (17-55 f/2.

, ultra-wide (10-22), and ultrazoom (18-200) covered, and that's not counting the 17-85. All at pretty reasonable prices relative to quality; and the ones that could stand to be better (17-85, 18-200) probably wouldn't be appealing at the resulting higher prices. An updated 17-85 would cost the same as a 15-85, so why not just get that already. And an updated 18-200 would rival the price of the 24-105, etc...it'd have to get a pretty major IQ upgrade to be worth the price.
The travel superzoom segment is already well covered by sigma and tamron, who can sell to many more potential users of different systems.
And then there is this. Making an 18-270/18-300 that is good is not easy, it's usually heavy and bulky and not as desired as people think. So, if you can make one for 3-4 mounts, maybe you can make money...but just making it for EF, that's a tougher sell. Better to release a bunch of decent travel-zoom P+S cameras instead.
I'd much rather see Canon release an EF-S 30mm or EF-S 12-14mm prime before I'd want them going super-zoom. Nikon and Sigma show that the 30/35mm aps-c specific lens can sell; and Samyang seems poised to show that a 10mm prime can sell as well