Small cells negate the "limited amount of spectrum" argument. It's a financial + logistical + political/regulatory limitation, not a technical one.
Technology will eventually advance to the point that the financial consideration is less important. We're already working with beam-forming -- a technology that's existed for decades, in radar applications -- for instance. Wireless is the future, no matter what the naysayers think, and if you're still thinking of "spectrum" as the limiting factor you're behind the curve. Makes me think of the folks who deploy IPv6 for the first time and start worrying about the "waste" of addresses.