When I was an undergrad at the University of Pennsylvania in the nineties, there was a joint project between Penn and MIT to build the world’s fastest, long-distance Internet connection between Philadelphia and Boston. In 1993 the project was successful, creating the first long-distance 2 Gb/s Internet link. At the time, it was the fastest link on the Internet. I remember doing some back-of-the-envelope calculations and thinking that we were getting near the limit of what a single computer could handle.

Fast-forward 18 years. A few weeks ago, Orange and Alcatel Lucent turned on a 17.6Tbps link between Paris and Lyon. That’s “T” as in 17.6 trillion bits per second, making the Paris-Lyon link about 8,000 times faster than the UPenn link in a span of 18 years. Do some quick math, and you’ll find that we’ve accrued about 65% interest on speed (compounded yearly).

I bring this up because that 8,000 to 1 ratio also holds true when you look at the 100Kbps ISDN modems that were state-of-the-art when I was in college, compared to the 1 Gbps connections that Google Fibre is now delivering to folks in Kansas City.

Today’s fastest consumer Internet is basically the same speed as the world’s fastest long-distance Internet link 18 years ago. At this rate, we’re trending towards having 17.6 Tbps Internet connections in homes by 2031. Today, it’s clear that processing power will keep up with ever-increasing Internet connection speeds. So, we’ll all have 17.6 Tbps Internet connections in the 2030s, and it won’t seem fast enough. The real question is: what will we be doing with all of that speed?

Alex Gizis is CEO of Connectify. Follow him on Twitter @agizis and visit www.connectify.me