NYC-Chicago was the most important link. They went back-and-forth between using cables and using microwaves. Both are slowed down by the path length and by the number of repeaters required, straightening out the path is expensive but allows fewer repeaters. Like the Pony Express, sometimes just when you think you are going to make a profit, somebody else puts in something faster.
Between NYC and London, there was noticeable echo on the original line – a noticeable fraction of a second. That was replaced by satellite communications, which was so much worse that digital communications all preferred the old cables. That was the path length from here, to up in the sky, to back down again. Optic cables were much better, and much faster.
All of that was much better than shortwave radio, which was flakey at the best of times. but, the same technology that has made other comms possible also can make digital shortwave radio better. You can use broadband spread-spectrum and frequency hopping. And the nature of high-frequency trading is that you can simply turn it off and exit the market when you don’t have communications: you miss profit, but you aren’t holding a position.
So about 20 years ago people started working secretly on High Frequency Trading with shortwave between NYC and Chicago, and 10 years ago they were trying to do the same between NYC and London. How much success they’ve had, I don’t know, because, unlike laying a new cable or placing new microwave towers, there is nothing to see. A little bit of information from the FCC, and from strange hobbyists who look at antennas instead of trains. I don’t think it’s gone mainstream, but I’m out of touch.
(It’s not random for me – I worked in telecom and in finance)