That's still not a bad investment vs other investments you can make.
It totally depends on the increase in power of the total network. If the total hashing power increases by 10 or 100-fold a year (20 fold increase in the last year), many miners will never break even. And if they do break even if the price of bitcoin spikes, it's still quite possible it would have been a better to put money spent on rigs into bitcoins at the start of the period instead.
The real question people looking to buy an ASIC should ask themselves: "Do I think this will generate more bitcoins over its lifespan than I would by buying bitcoins right now with the money I would spend on equipment and power costs?" Because if BTC is at $100 and a rig is $5000 and you don't think you'll ever generate 50 BTC with it (plus enough to cover power), there's not much point, even if you think 50 BTC might eventually exceed $5000 in value.
