ViaSat1 has an overall capacity of 140Gbit/sec and that was launched in 2011. I'd imagine that higher capacities are possible (though considering they are very low weight, that may restrict the amount of beams it can push out).
However, the continual problem with satellite internet is that the satellites have a life expectancy of 15 years or so. While 140gbit/sec+ is probably enough now, it certainly won't be much in a few years time (when you have to consider that the bandwidth is shared across an awful lot of potential users).
Edit: I'd also add that while the proposed market is for areas without current internet connectivity, people do need to realise that a lot of Africa has very good 2G/3G coverage and some areas have very good LTE coverage. The problem historically has been backhaul capacity
a) from the towers back b) from the core network back to the internet
a) is still an issue as in Africa most if not all backhaul from the towers is provided via microwave relay and daisy chaining them, which causes capacity problems (most of the West has fibre or similar connecting each tower)
b) is slowly being improved with much better fibre runs around Africa.
Considering that most very basic smartphones will have LTE chipsets by the time this system is up, it may be more cost effective to use LTE than this system. This system is likely to require expensive ground terminals that need to be fixed in place.