I think also part of a long term strategy, in that AM in the past has demonstrated that they are conscious about not wanting to disrupt the network by owning a large share of it (say 40 or 50%). Franchising and cloud mining are avenues to manufacture chips and miners for more than their share of the network, and still seeing profits trickle back to AM from those.
Franchising would work this way, but cloud mining would not since the hashpower owners are not the ones controlling what transactions are included in a block. It's essentially like mining to a pool you can never switch away from.
Not that I think AM will ever come close to 40%
You are correct here, although it seems to me from the description on the website that the cloudmining service itself is a franchise... I don't know anything for certain of course.
I don't know if AM will come close to 40% again or not, but they have one thing working for them in that regard; That they actually plan how that might happen, and make a roadmap accordingly. Pretty much all of the Asic failures didn't have a plan beyond continually scaling up what worked last time, without ever revisiting the idea to make sure it's valid in the current environment. Look at BFL for a stunning example of this, I mean, if you ignore their failure to deliver, you see that they didn't really have a chance anyways, as they were still planning to sell fancy overpriced aluminum boxes to hobbyists, as if we were still in their FPGA glory days. You need a plan as much as you need to act on it.