Well it's a couple of things.

First off, a wireless transmission speed of 120Gbps sounds really impressive but remember from the Shannon-Hartley theorem that the maximum channel capacity is just a function of bandwidth and SNR. This means that you can get an arbitrarily high transmission speed by increasing bandwidth to an obscene amount and/or by increasing SNR (by transmitting at an obscenely high transmission power).

In the paper they say that the transmit power was 15 dBm which is a normal transmit power for WiFi, so it's the 40GHz bandwidth that's doing the heavy lifting in allowing that data rate.

The second thing is that WiFi 6 (for example) uses 1.2 GHz of bandwidth in the 6GHz range, divided into seven non-overlapping 160MHz channels. WiFi 5 uses about nine 80MHz channels in the 5GHz range, and so on. So if you want to use the technology demonstrated in the paper for WiFi (as the headline of the article is suggesting) then you'd need a bunch of 40GHz channels in the higher ~200-300 GHz range which would be in the very high microwave range, bordering on far infra-red!

If you want to imagine how useful that would be, just think about how useful your infra-red TV remote is. You would only be able to do line-of-sight point-to-point links at that frequency.

IR point-to-point links already exist, and the silicon they invented for this paper is impressive, but the hype around it being a possible future WiFi standard doesn't really hold up to basic inspection.