In case you missed it, RCR Wireless ran a webinar session last month under the heading ‘Private 5G for IoT – plotting timelines, defining applications, and making bets’. The session is available on-demand, and also attached to a full editorial report, out in a couple of weeks. But ahead of that, and just to keep the momentum, here is a snapshot of the webinar dialogue, which brought together Leo Gergs from market advisory firm ABI Research, Jagadeesh Dantuluri from test and measurement outfit Keysight Technologies, and Kathiravan Kandasamy from carrier services provider Syniverse; all of them are squarely focused on the private networks game.
This summary article, to be extended in the editorial report, is in three parts, as per the headline description; we start with ‘plotting timelines’, and hand directly to Gergs, with a forecast about how the private 5G sales market plays out in the period to 2030 – to the tune of $50 billion, eventually, and significantly more with non-5G cellular networking on top. He comments: “Our analysis says that the overall market for private 5G will be a bit more than $50 billion by 2030. If you add private 4G/LTE, then it is roughly twice as much. By 2030/31, we will reach a turning point, where private 5G takes over from 4G as the dominant cellular-based enterprise networking technology.”
So, there it is: private cellular will be an LTE market for another five years at least; the road for private 5G is a long and winding one, yet. At the same time, of course, the incline is rapidly steepening, to get to that changeover point. The question goes to Dantuluri, also acting chairman for one of the 5G-ACIA working groups, looking at IT/OT crossover in industrial 5G systems: does that $50 billion / 2030 stat tell us how far private LTE has come, or how far private 5G has to go? He responds: “Most deployments, today, are on LTE; but 5G is growing much faster. So there is a flipping point sometime next year where most of the new networks will be on 5G rather than on LTE.”
Which is another marker in the ground – that from 2024-“sometime”, 5G starts to loom-larger in the LTE rearview. Dantuluri says: “[The industrial sector turns on] a 10-year lifecycle; [it doesn’t] want to go 5G [yet] because the cost of 5G devices is too high. LTE performance is good enough for their use cases [now]. But… [Release 16] devices and networks are coming up pretty fast, [and they] are the central theme of the whole industrial IoT movement – and they are needed, and LTE will [eventually] not be sufficient… By 2030, and maybe a little earlier, 5G networks [will be] as seamless as Wi-Fi today, and the cost [will have] decreased, and then we will see what is possible.”
So, this 5G-future is around the corner, but also somewhat out-of-sight; meanwhile, and still, even just for another 12 months, LTE is tearing up the road, with clearer navigation and a cheerier payload. Gergs suggests the big promise of private 5G (for industrial IoT) was a scam – at least until now, at least for paying customers. Release-16 systems will deliver (some of) the industrial-grade networking pyrotechnics written into 3GPP’s ultra-reliable low-latency (URLLC) and massive machine-type (mMTC) development streams. But there just ain’t much in the way of Release 16 chips and hardly anything at all in the way of Release 16 devices, as yet, to attach to Release 16 5G networks.
Gergs says: “The key bottleneck continues to be Release 16 and 17 industrial-grade chipsets and devices. They are beginning to emerge. But it’s important to stress [that their absence] is in the context of industrial-grade hardware; in the consumer market, for consumer devices, we already have 16-capable chipsets, mainly from Qualcomm. But we are not there yet with industrial-grade devices with enterprise features. Towards the end of last year, the talk was that these would start to be available in Q4 2023, reaching scale in the first quarter of 2024.” So we have another staging post, then: the start of 2024, realistically, for industrial-grade Release 16 IoT units to come available.
He goes further; the timeline for the arrival of their Release 17 equivalents, to push the industrial IoT journey even further along, is pegged for the end of 2024, about 12 months later – presumably for initial sampling, and not a full-blown commercial rush of new-fangled IoT hardware. He says: “We’re pretty much there. Release 17, which standardizes a lot of the important features of RedCap, for example, was frozen in March last year, and a typical cycle would see chipsets follow 18 months after. But we’ve seen with Covid, and with chip firms more focused on consumer features, that this does not necessarily hold. So I would put it at two years – as the offset, not a delay.”
Which is how we get to Q4 2024, for higher-fidelity 5G-for-IoT. But for the next 12 months, at least, private cellular is all mobile broadband (MBB), even when it is dressed up as go-faster LTE on 5G. Gergs references time-sensitive networking (TSN) as the other big IoT promise that cellular is yet to deliver. “The traditional value proposition was around URLLC, TSN, and mMTC – all of which will enable the automation and digitization of operational technology (OT) and OT data. But we are seeing a shift in the value proposition, right now, as the market considers Release 15 capabilities in private cellular. So we are looking at what can be done with eMBB capabilities.”
This is important, to get a handle on the private cellular market today. Because while 5G struggles to meet its various commitments, to support IoT use-cases that will spur the Industry 4.0 movement to make new gains, everything else comes down to money. The question is: why buy an expensive 5G network when you can do the same work with a cheap LTE network? In some cases the question is redundant, just because the business case for 5G, over LTE or Wi-Fi, does hold up. An example, raised repeatedly in the webinar, is to connect large-scale outdoor utility networks or big port areas, or other major industrial campuses. Just because the 5G-cell signal / capacity calculation is better.
Gergs says: “The market is mainly driven by enterprises with large outdoor areas where private cellular is just a much more efficient way to provide coverage – over a wider area with less infrastructure, compared to Wi-Fi. We are talking about a factor of 10 or sometimes even more than that. There is a lot of demand coming from energy generation, for example, and from logistics – from ports and airports, and in some manufacturing as well.” But otherwise, for indoor IoT applications, the business case for private 5G tends to fall down; many of these mMTC cases are already served by a mishmash of non-cellular BLE, LoRaWAN, and Wi-Fi networks, plus others.
Which is why the telecoms market desperately needs reduced-capability (RedCap) 5G to come quick, and to stick for Industry 4.0 users. Gergs again: “With regards to IoT, the competition is fierce. 5G has to compete with a lot of IoT technologies, mainly on price. Because – let’s face it – what drives the decision to invest in technology is the price. And 5G has to come down in price in the short term. Because as an umbrella technology, which combines lots of different use cases, IoT for mobility and positioning is not there yet. Which has to do with this bottleneck of Release 16, 17, even 18, when you talk about centimetre-level positioning.”
He closes this first section: “So at this point in time, 5G has to compete more or less directly with well established IoT technologies, mainly on price. Which is why all hopes are pinned on RedCap at the moment. Which is designed for use cases that lie somewhere between traditional 5G use cases that are in terms of high availability and reliability, and URLLC features, but require more licensed spectrum and more deterministic networking than can be offered by current LTE-based LPWA technologies like NB-IoT and LTE-M.”
We will handover to the others for parts two (defining applications) and three (making bets), but Gergs’ input, largely, sets the whole scene: private 5G will take 12 months to get real, 24 months to get really real, and six years to be considered properly mature.
