For years it’s been impossible to go to any industry forum without meeting a few folks who predict that residential broadband will go wireless. This buzz has accelerated with the exaggerated claims that fast 5G broadband is right around the corner. I’ve seen even more talk about this due to a recent Pew poll that shows that the number of people that only use their cellphones for data has climbed significantly over the last few years—I’m going to discuss that poll in another upcoming blog.

The question I’m asking today is if it’s possible that most residential broadband usage in the country can go wireless. Like I usually do, I looked around the web to try to define the current aggregate amount of landline and cellular data currently being used in the US. It’s a slippery number to get a grasp of for a number of reasons, not the least being that broadband usage is growing rapidly for both cellphones and landline connections. It looks like landline data usage per household is still doubling about every three years; it looks like cellphone data usage is doubling every two years.

OpenVault recently reported that the average monthly household broadband usage has grown to 273.5 gigabytes for the first quarter of this year, up from 215.4 gigabytes a year earlier in 2018—a growth rate of 27% which almost exactly doubles usage in three years if sustained.

There are currently a little more than 127 million households, and the FCC says that around 85% of all households have broadband. Extrapolating that all out means that US landline networks in aggregate carried almost 30 exabytes of broadband for households monthly in the first quarter of this year. (An exabyte is 1 million terabytes, or 1 billion gigabytes).

I’ve seen a few recent statistics that say that about 77% of Americans now have a smartphone, up from 67% in 2017. Recent statistics from several sources say that the average data usage per smartphone is now over 4 gigabytes per month, with buyers of ‘unlimited’ data plans averaging more than 6 gigabytes per month and others still down closer to 1 gigabyte per month. With a current population around 329 million and using an average of 4 gigabytes per month per residential phone, the cellular networks are currently carrying about 1 exabyte of residential broadband per month.

If we extrapolate forward six years, assuming keeping the existing growth rate for each kind of broadband, we can predict that total monthly US residential broadband usage will be something like the table below. Note that these figures exclude business broadband usage.

Today the landline residential broadband networks are carrying 29 exabytes more of data per month than cellular. Within six years, that difference grows to 93 exabytes. There is no reasonable path forward that will have cellular data usage overtake residential usage in our lifetime.

The next issue to address is the overall capacity of the cellular network. The engineers at the cellular networks are likely cringing at the possibility of having to carry 6 exabytes of cellular data per month in six years—a 600% increase over today. The cellular companies are going to be increasing data capacity in three ways—adding small cells, adding more mid-range spectrum, and adding 5G efficiency captured mostly through frequency slicing. It’s going to take all of those upgrades just to keep up with the growth in the above chart.

There are those who say that the way the cellular companies will handle future growth is through the millimeter wave spectrum. However, that technology will require a fiber-fed small cell site near to every home. We really need to stop referring to millimeter wave spectrum as 5G wireless and instead call it what it is—fiber-to-the-curb. When thought of that way, it’s easy to realize that there are no carriers likely to make the investment to deploy that much fiber along every residential street in America. Wireless 5G fiber-to-the-curb is not coming to most neighborhoods. The bottom line is that the world is not going to go wireless, and anybody saying so is engaging in hyperbole and not reality.