Ericsson has published the latest edition of its Mobility Report, charting the growth of the mobile market, the size of the IoT contingent within it, and its expectations for 5G. While much of the report deals with the volume of data and the demographics of the markets, we are of course most interested in the IoT side of things.
It will come as no surprise to see that Ericsson points to 5G as key to facilitating smart cities, IIoT, AR, autonomous transport, and digital health. Highlighting the current state of the market, Ericsson has three examples with its own customers, namely MTS (Russia), Telstra (Australia), and Turkcell (Turkey). The full report is available here.
In Q1 2019, Ericsson says there were 7.9bn mobile subscriptions globally, up 44mn in these three months. Mobile broadband subs grew 15% year on year, reaching 6bn or 76% of total mobile subs. LTE subscriptions grew 160mn in the quarter, reaching 3.7bn or 47% of all mobile subs. GSM/EDGE-only subscriptions fell 80mn in the quarter. Smartphones account for 60% of all mobile phone subscriptions, and mobile subscription penetration is above 100% in every territory except for Africa and India.
Ericsson says that over 10mn 5G subscriptions are projected to be active by the end of 2019, and adds that uptake is expected to be significantly faster than LTE. To this end, it projects 1.9bn 5G subs by the end of 2024, accounting for 20% of total mobile subscriptions (8.3bn). LTE will peak in 2022, at around 5.3bn subs, and based on the graph, Ericsson projects that there will still be around 400mn 2G subs, and around 1.2bn 3G.
This brings us to the IoT section, where we need to define terms. Ericsson uses the Cellular IoT umbrella term as a catch-all for Legacy (2G and 3G), Massive IoT (LTE Cat-M and Cat-NB), and what it calls Broadband and Critical IoT (4G and 5G).
Based on its graph, Ericsson says that Legacy connections are going to grow from around 800mn in 2019 to hit 1bn in 2022, at which point they stay flat until 2024. Massive IoT grows from around 200mn in 2019 to around 1.7bn in 2024, with Broadband and Critical growing from around 500mn in 2019 to 1.5bn in 2024.
Ericsson stresses that Cat-NB and Cat-M are going to be able to fully co-exist in the same spectrum bands as 5G-NR, which is a similar message being pushed by Sequans and u-blox, in their recent announcements.
As for scale, Ericsson says that there are around 1bn Cellular IoT connections in 2018, and that there will be 4.1bn by 2024 (CAGR of 27%). It says that “Wide-Area IoT” connections are going to grow from 1.4bn to 4.4bn, but that “Short-range IoT” connections will grow from 9.3bn to 17.8bn in the period.
Unfortunately, Ericsson is still not giving up on the idea that 5G can replace wired connections in industrial automation settings. We maintain that any operational manager worth their salt will laugh the wireless salesperson out of the room when they suggest ripping out ethernet runs and replacing it with 5G, regardless of what the 5G Alliance for Connected Industries and Automation (5G-ACIA) working group might be developing.
Ericsson also touches on 5G coverage in the report, saying that 5G can reach 65% population by 2024. 4G grew by 10% population coverage, reaching around 75% at the end of 2018 and expected to reach 90% by 2024. 3G is around 95% now, and is expected to remain that way through the period. Tellingly, there’s no mention of geographic coverage, which is the stick we usually use to beat the L-LPWAN crowd with.
Interestingly, Ericsson also looks at the demand for private cellular networks in the report, saying that the two main drivers here are Land Mobile Radio (LMR) modernization, for things like public safety and natural resource extraction, and Industrial Digitalization.
It is notable that there is no mention of unlicensed spectrum when Ericsson says that “a key challenge for these deployments is how to access suitable spectrum when using cellular technologies for private networks. There are three main alternatives: via service providers, direct allocation by regulators to industries, or by using shared licensed spectrum with local allocation. Which option is best will depend on the conditions in each market.”
Another interesting part of the report is that where Ericsson outlines how a computer vision system, attached to a drone, is able to enhance tower inspections and reduce the risk for workers having to climb up and down more frequently. A computer-vision application running on a workers mobile device is apparently powerful enough to power this function, but one does wonder if a human would be just as able to spot problems with the tower from the video feed.