4G used clever ideas to get much better practical spectrum efficiency than people thought was possible... At the cost of very expensive infrastructure. The biggest selling point to carriers was standalone 5G networks that were cheaper to build but paradoxically carriers are prone to frame 5G as a feature customers pay more for.
The area where wireless networks fall short is coverage. Obviously many homes and businesses are difficult to cover with wireless technology. Also many "IoT" and public service situations need wide coverage, and a network that works in 92% of places is "close but not cigar." There are few obvious answers rather than "lots of fiber and very cheap small cells" and "satellite technology" but it's what it takes to address wireless use cases that currently aren't addressed.
e.g. Notably several major carriers in India skipped 3G entirely and went to 4G directly because it was cheaper per subscriber. Which implies less expensive equipment per unit performance?
The way I remember it in the US was in the 3G age you had good cell service in cities like Rochester, NY which have moderate populations and gentle hills but it was spotty in dense cities with crazy geography and/or urban canyons. [1] Like I used to stay at a motel two blocks from Hollywood and Vine and my phone barely worked. Once 4G was rolled out, coverage got good in Nielsen "A" markets.
[1] ... and there was that summer my son and I tried to visit as many state parks in upstate NY as we could and I had phones on both the AT&T and Verizon networks that had terrible coverage away from certain big roads.
The area where wireless networks fall short is coverage. Obviously many homes and businesses are difficult to cover with wireless technology. Also many "IoT" and public service situations need wide coverage, and a network that works in 92% of places is "close but not cigar." There are few obvious answers rather than "lots of fiber and very cheap small cells" and "satellite technology" but it's what it takes to address wireless use cases that currently aren't addressed.