Interesting idea, can't really assess it with out any solid numbers though. The economist makes the comparison with greenhouse farming, which is (as the economist says) very energy inefficient, also currently it produces lots of agricultural effluent, but with hydroponics you wouldnt get this. The numbers we'd need are inputs, outputs and construction costs. You can argue that it may one day become cost effective, but imo only if the construction and energy costs fall, as a higher food price would not be a good thing.
If we are worried out feeding a massive world population there is one very easy and very unpopular way of doing it. Slash world meat production by 50% (possibly less, ive just pulled that stat out of thin air but it makes the point). Meat, in particular beef is the most inefficient form of food production in terms of land, energy etc. Most efficient i think is soya beans, or some other pulses. Wheat and rice are ok. Its also worth saying that food security is less an issue of global production and more an issue of distribution. The world produces numerically enough food currently, and could produce more, most first world countries reduced production through the 1990's. The issue is over production in the 1st world, under production else where. Having said this, i suspect an increase to 15b would exhaust current production by a clear margin. I dont think growth will exceed 10b tbh, im a malthusian, possative feed back will kick in harder sooner or later.