Shannon showed that the maximum capacity of a channel is given by:
$$C = B \log_2\Bigg(1+\frac{S}{N}\Bigg) $$
Where \$C\$ is the capacity in bit/s, \$B\$ is the bandwidth in Hertz, \$S\$ is the signal power, and \$N\$ is the noise power. In most case we are limited by thermal noise, and in that case we can only increase SNR by increasing signal power. This is possible with better and more efficient antennas/filters, or starting out with more power in the first place.
The thing is that if our signal power is much, much higher than the noise power, we might be limited by the noise power that is actually transmitted by the transmitter. This noise can be due to the noise and distortion in the mixers and amplifiers, phase noise in the synthesizers, etc.
So getting higher signal power will help, but at some point it gets harder an harder. In addition, the SNR part is in the \$\log_2 \$ term, in other words, you get diminishing returns. Usually, it is easier to just increase the bandwidth instead - which is (one of the reasons) why people are so interested in the 60 GHz and other millimeter wave bands for communications.