-- Keep in mind that I know nothing of nothing.--
--- Is the sub-millisecond from repeater to repeater or from origin to final destination that's [n] repeaters away? Won't each repeater introduce its own latency?
What causes the 4G's >50ms response time?
So let's say you're in Scotland, the packet would enter the mobile operator's network, then go to London, where they've bridged the network to the internet.
So there's an extra there and back again.
For example you can have a 5G core over LTE.
Your 4G results also seem quite bad, my personal test right now was more around 18ms broadband vs 31ms LTE. Google scholar results (https://scholar.google.de/scholar?hl=de&as_sdt=0%2C5&q=LTE+l...) I looked into seemed to suggest LTE is capable of ~20ms latency results, while (http://wirelessone.news/10-r/1007-lte-latency-today-9-ms-dow...) suggests that LTE-A (3gpp release 13) is capable of 14ms latency.
5G has the potential to provide throughput and latency that is comparable to a fixed broadband connection. In reasonably competitive markets (i.e. not the US), that's A Big Deal. Latency is a particularly acute issue in many applications; 4G generally adds about 50ms in the best case scenario, but 5G can easily provide sub-millisecond latency. Imagine a near-future where it simply doesn't matter whether you're on WiFi or cellular, because they both provide the same experience.