DOCSIS3.1 over old copper coax cable, with enough RF channel bonding, can also do 300 to 900 Mbps downstream, again slightly oversubscribed. With much less upstream because the RF channels are usually allocated asymmetrically at a ratio of 10:1 or thereabouts.
This is by no means going to hurt Comcast. Comcast (and similar companies like Wave, RCN, Charter, Shaw, etc) are building DOCSIS3.1 and also overbuilding their own networks with fiber. They have the established right of way through major cities, both on pole and underground, and can incrementally build a FTTN/FTTdP architecture with existing copper, and will eventually overbuild everything with singlemode fiber to the customer.
What this will be revolutionary for is small towns, remote/rural areas. In the US as an example the market for ISP last mile connections that is presently served by either WISPs, or in places where no WISP exists, two-way Ku/Ka-band VSAT terminals (HughesNet, ViaSat, etc). The capacity constrains on traditional geostationary consumer grade VSAT services result in terrible things (TDMA oversubscription that pushes latency to 1150ms during peak hours), and things like 20GB/month transfer quotas. That's for a typical $100/month, 24 or 36-month term with the latest generation of Viasat's services.
For polar regions this will be a big deal because they are presently almost entirely dependent on geostationary C and Ku-band satellite capacity.
On ships: Very relevant and has the potential to capture a ton of market from Inmarsat's I-4 and I-5 series satellites and tech, and Iridium's upcoming maritime products. With a sufficient number (thousands!) of polar orbit satellites it can cover mid ocean areas very effectively.