There are physical limitations.
With high order modulation schemes you need about 3dB Carrier/Interference(C/I) ratio for every bit of throughput per hertz of spectrum. This comes out of information theory - It is a physical limit.
LTE operates within about 1dB of this physical limit plus around a 20% overhead for channel signalling. You can use signalling redundancy to tradeoff throughput for C/I. The eye-watering throughput claims of operators simply do not apply once you have multiple users on more than one LTE site in operation because they interfere with each other. In practice on a good day you might see around 20mbit downlink performance per cell sector on a 10MHz LTE channel and that is using MIMO. This figure will slowly increase in the future with higher order MIMO and also become more reliable with beam forming techniques.
You cannot reduce the noise in the channel by fancy coding schemes only by raising your transmit power and QAM based systems use extremely linear transmitters to maintain orthogonality. If the new proposed system dispenses with the linearity requirement you can improve transmitter power efficiency and raise the power with existing devices thereby giving you an improved C/I and more throughput. This is providing the telecommunications regulatory authority lets you increase the power. Cellphone transmit powers are limited to stop you absorbing too much signal with your head.
In any case most of the time a connection is 'interference limited' in that other transmissions impact the theoretical link C/I and reduce the throughput. If you have a short range connection then you may benefit with the new system but I suspect the improvements will be moderate rather than dramatic.
Andy