What is ‘Forecastable’ About 5G?
A frequently asked question of industry analysts is “how big is the 5G market going to be?”. The real answer is that although it might be possible to forecast the size of certain aspects of 5G, it is nearly impossible to forecast the 5G market as an aggregate entity. Instead, I’d argue that there are several discrete, and not very related, elements of what might constitute a 5G forecast. Hopefully the thought process below will help those who want to gauge the progress and dimension of this exciting new phase of wireless and broadband.
First off, over the next couple of years, there will not be a distinct 5G service, in the conventional way we would think about it. It will be a few “islands of 5G” in a sea of 4G LTE – where a relatively limited subset of cells in a certain market will be equipped with 5G radios. It will look more like roaming onto a 5G hotspot (like Wi-Fi) than a seamless service – even from those who say they’ve deployed ‘mobile 5G’. Some operators might even obfuscate real 5G with services that might look like 5G, or elements of the LTE roadmap that boast 5G-esque characteristics or performance. Full throttled gigabit LTE service will feel as good as early 5G, as we’ve seen from some of the marketing, and even branding, such as AT&T’s 5G Evolution.
The equipment deployment aspect of 5G is going to be especially tricky to count. Many operators are adding small cells and densifying their networks in order support more capacity for LTE, and also to be ready for the higher bands that will be used for 5G, such as 3.5 GHz or mmWave. While 5G might be a driver for these deployments, or a beneficiary of them down the line, they still fall in the 4G bucket.
So, one way to look at this might be measure the percentage of cell sites in a given area that have been equipped with or upgraded to 5G. Keep in mind that many of the sites that have been shipped over the past 2-3 years are 5G ready, meaning they can be upgraded to 5G with a software update.
Second, some of the 5G services will be based on specific business cases, but will be counted outside of some of the metrics we typically use to count things in the mobile industry. Case in point is Verizon’s soon to be launched fixed wireless access (FWA) service. Here, 5G might be the access mechanism, but really this sits in Verizon’s fixed broadband bucket, as in Fios or DSL. No Verizon Wireless customer will be able to buy this service unless they want to use it for broadband in their home.
This gets even tougher with other markets that will drive the 5G business case. The IoT aspect is a good example. One of the important capabilities of 5G is that it can handle millions of connected devices simultaneously. But it could well be that these are many millions of low power devices consuming very little data. Not the ‘faster, higher, stronger’ flavor of what we normally consider in a wireless network upgrade. Or, take connected cars. Vehicles using V2X will employ 5G for some aspect of a vehicle’s overall communications, but this will be a challenge to actually count.
The most countable – and hence most forecastable – aspects of 5G will come from the equipment and device sides. It will be easiest in greenfield implementations, such as in mmWave, where all the radios are 5G by default. It is possible to forecast and count 5G NR, even though they might not be the same as a 4G radio in the physical sense of the word. It will also be possible to forecast 5G devices that can talk to a 5G radio. Initially it will be 5G bricks and dongles, but in 2019, we’ll see some early 5G phones. There will also be a much larger number and variety of devices that are 5G enabled, meaning they contain a chipset that can talk to a 5G radio. Another new ‘category’ will be the CPE for fixed wireless, both at the site as well as the little box that sits outside the window.
There is also a slew of technologies and products that will deliver some of the capabilities that will make 5G distinct. This includes advanced antennas, the techniques behind advanced beam forming, and network slicing. These are all parts of the 5G ‘market’, but difficult to isolate, as in “the market for network slicing is X”. It is more that the capability for network slicing is one of the key features of 5G, and might help open up new frontiers in the enterprise mobile market.
Those who have read this far might argue that much of my argument could well apply to what we saw with 4G LTE. But I’d argue that this is different, and more complex. 5G will be rolled out on a more staggered basis, on timetables that vary significantly from one country (or city) to the next, and only sometimes on purpose-built spectrum. It will be spread across a much greater variety of use cases/application types and on a much broader suite of devices, not all of which might use 5G in the conventional sense. Actually, a minority of 5G devices will be phones as we know them today.
5G is going to be something very significant over the next 10 years, but it will be some time before we can get our arms around what actually constitutes the “5G Market”.