5G Availability Data
There has been much discussion in recent days of an OpenSignal report[1] that shows the percentage of time 5G is available to users in the USA is 8% for Verizon and 12% for AT&T (but an amazing 68% for T-Mobile). Given all the expenditure on 5G and all the hype, the fact that many users can only access it around 10% of the time sounds shocking.
In fact, the low numbers in the US result are quite typical. OpenSignal report that many countries have 5G availability figures around 10%. T-Mobile is very much an outlier. Even in the “poster-child” 5G deployment of South Korea, availability ranges from 35%-45%, but in, eg Canada it varies from 8%-10%, in the UK it is 10% across all MNOs, in Germany 9% - 16% and so on. Saudi Arabia is better at 18%-20% but has long had policies to promote 5G deployment. OpenSignal only lists[2] six operators globally that have more than 30% 5G availability.
Operators claim much higher levels of 5G population coverage. But the operator numbers are typically associated with achieving the minimum signal level outdoors. Many users spend much of their time indoors and higher frequencies, such as 3.5GHz, suffer relatively high building penetration losses. And networks may move users onto 4G where the 5G signal is weak but there is good 4G signal. The OpenSignal numbers have the advantage of showing what users experience, taking all factors into account, not what computers predict will be areas where there is 5G signal.
Does this matter? Broadly not. Users in most countries are seeing data rates far higher than the 10Mbits/s or so beyond which users notice no benefits. 4G availability is generally very good – 99% of the time for example in the US. Users have no need of greater 5G availability.
As a result, this level of availability is likely logical for MNOs who use 5G more to enhance capacity than to deliver new services. Spending more might increase the percentage 5G availability but the consumer experience would be unchanged and hence revenues would not increase. As I showed in “The end of telecoms history”, data usage will soon plateau and so the need to deploy further 5G for capacity reasons will be limited.
Because of these factors, the current typical 5G availability of 10-20%, which has taken five years of 5G deployment to reach, is unlikely to change much. Or if it does, it will be because sub 1GHz 5G is deployed, which brings no benefits over 4G. It seems likely that most will never see 5G availability for more than a quarter of the time. As I mentioned earlier, they likely will not care.
But there are some important implications.
Firstly, the incentive to develop any apps that require 5G performance (eg very high data rates or very low latency) is low. Customers are not going to be happy with an app that works only 10% of the time. And with no apps that need 5G the incentive to deploy 5G further is low, feeding a vicious circle. The incentive to deploy 5G SA is likewise small – why deploy solutions like network slicing when 90% of the time users won’t be connected to a network that can be sliced?
Secondly, the availability is heavily influenced by the frequency band. 3.5GHz has materially worse propagation than bands used for prior generations. And yet 6G proponents are looking at even higher bands in the region 7-14GHz. This could result in 6G availability remaining in single figure percentages for its entire lifetime.
Of course, this was all entirely predictable – see “The 5G Myth” published in 2016 that suggested 5G availability would be minimal. The surprise is that it is as high as 30% in a few isolated cases.
Programmable Telecoms / Communications Expert
7 个月Why is the industry pushing 5G Network APIs when 5G availability is so low? No developer would build for a capability that is only available 10%? Lat's park the P5G discussion as that can be solved today, like it is for P4G today.
Builder and Consultant on Open vRAN, Small Cell and EdgeAI Networks
7 个月I don't see any UK data for 10% in the links that you sent, please could you share the reference? I'm curious that you discuss 5G in isolation, when from a deployment perspective, it's not. If people are satisfied with a 4G connection, it's because 2/3 of the site capacity upgrades associated with 5G NSA deployment is on 4G. Single-band 4G 2T2R RRUs replaced with multi-band DSS 4T4R Radios. If there had not been a 5G deployment there would have needed to have been a 4G+ deployment. The lack of suitable 5G small cell upgrades for 4G small cell sites at launch is another factor. The small cell site needs to deliver RF dominance and will therefore serve the user even with 4G+ in the presence of a 5G Macro signal. As you say, the user is happy with the 100Mbps+ these sites deliver, but they will have a 5G capacity uplift delivered in most cases, almost every DSIT ONE Trial uses 4G&5G densification as a Neutral Host.
The real problem is that no customer cares. Beyond the NHS there cannot be many spends in the trillion range where the users of the service are not sure what has changed...
M&A Pre-Investment Due Diligence; Market Development; Technology Strategy
7 个月Carrier aggregation makes it possible to get sufficient high speed using LTE. 5G as deployed today is only a massive MIMO technology to allow mid-band spectrum deployments, which LTE could not do efficiently... The industry cannot help but create inflated expectations... Fundamentally, 5G is nothing more than another frequency overlay...