Discussing LTE-M and the Future of Cellular LPWA

On paper, the Narrowband Internet of Things (NB-IoT) has been more successful than LTE-M. Nearly 105 million NB-IoT modules were shipped worldwide in 2021, versus just over 23 million LTE-M modules shipped in the same year. But it is no secret that China is responsible for consuming the majority of this NB-IoT bounty, with an estimated 66% share in 2021 that equates to 70 million units. Similarly, it has been Chinese module vendors that have sold most of all NB-IoT modules, 96% of those shipped globally in 2021, and it was Chinese semiconductor manufacturers that supplied the majority of all NB-IoT baseband modems shipped. China, however, is a captive market that is effectively inaccessible to foreign module vendors. To identify their opportunity, foreign vendors are only interested in understanding the cellular Low Power Wide Area (LPWA) market as divided into two principal segments: China, and the rest of the world. Non-Chinese module vendors do ship product to China, but not for consumption there. Instead, their modules are built into devices by Chinese manufacturers for international export.

With no LTE-M networks in China, there is no need for Chinese module vendors to make use of non-Chinese chipsets, unless they are also intended for devices destined for export. There is also no need for Chinese semiconductor manufacturers to offer dual-mode LTE-M and NB-IoT chipsets, as this would not be useful in targeting China’s NB-IoT-only demand. Dual-mode chipsets and modules are slightly more complicated in design and slightly more expensive—costs that, even if minor, would have some negative impact on the profit margin per unit sold, a margin that is already very low in dollar terms. The Average Selling Price (ASP) for NB-IoT modules is US$3.57 globally, and even lower domestically in China. Of the near-150 NB-IoT module models that ABI Research catalogues as having been produced by Chinese vendors during the last three years, a quarter used single-mode NB-IoT chipsets supplied by HiSilicon. HiSilicon is Huawei’s semiconductor subsidiary and exclusively manufactures NB-IoT-only designs, of which it currently has three with its most recent being the Boudica 200 model being launched in December 2020.

If we exclude China’s share of the cellular LPWA market, LTE-M and NB-IoT are become more closely placed, with 23.5 million and 35.5 million shipments in 2021, respectively. The market share between Chinese and foreign vendors becomes closer too, with almost 25% of international LPWA modules (Cat.M and NB-IoT) attributable to the former and 75% to the latter. But in the five years since the launch of the first NB-IoT network in October 2016, and more than five years after the launch of the first LTE-M network in March 2016, these remain low numbers. This is a fact that has disappointed semiconductor manufacturers, module manufacturers, and wireless carriers. Long-Term Evolution (LTE) Category One (Cat.1), however, has done well in that time. Internationally, while LTE-M’s footprint has been gradually growing, Cat.1 with 2G and 3G fallback has become the affordable lower power, lower cost mobile technology of choice for the IoT. It enjoys wide support, having been part of the 3GPP Release 8 back in 2009, and functions anywhere LTE has been rolled out. Although not conceived as an IoT technology, Cat.1 in effect is one, being almost exclusively used by IoT devices, as the humble sub-High Speed Packet Access (HSPA) bandwidth it provides is of no use for consumer mobile broadband applications.

Cat.1 has been hugely successful in China too, where without LTE-M it has by default become the only feasible option for higher bandwidth, always-on, and/or mobile-centric lower-cost IoT connectivity. There have been sudden surges in demand for Cat.1 for topical high-volume use cases for several years, notably for shared power banks in 2020 for shoppers to keep their mobile phones charged in malls as well as cloud speakers in 2021, as used by shopkeepers to receive voice confirmation of mobile payments received in busy smaller premises without needing to keep checking their phones. The myriad of use cases that could become the next big thing, including vehicle battery monitoring systems, to show battery charge as well as vehicle location, heavy truck exhaust monitoring for pollution regulation, Controller Area Network (CAN) bus data acquisition, and Electronic Control Unit (ECU)-locking systems for vehicle security, as well as the nebulous shared economy at large, for equipment hiring, tracing, and return. But as an accidental IoT hero and a very convenient stopgap, Cat.1 only has a limited window of opportunity, as LTE’s days are numbered.

What will emerge as the next successful cellular technology for the IoT? There has been discussion for several years about Cat.1 Bis, the single antenna version of Cat.1. But as an LTE technology, it has no greater longevity than Cat.1, with even fewer years remaining to build supply-side support and an end user base. Large western wireless carrier groups have shown a keen interest in Cat.1 Bis to finally fulfill the massive Machine-Type Connectivity (mMTC) promise NB-IoT and LTE-M seem to have failed to deliver. But assumptions about Cat.1 Bis being a sublime combination of low-cost LTE-M with the performance of Cat.1 may be unfounded, as Cat.1 Bis may only be destined for service in China, and any other countries that do not yet have LTE-M, or for applications that also require voice. Longevity and the ability to build globally saleable IoT devices with a single Stock Keeping Unit (SKU) matter. The need for the guaranteed continuity of IoT service availability in terms of geographic coverage, and in time with individual IoT devices needing to be functional for ten years, is a constant as old as the machine-to-machine (M2M) connectivity itself.

Cat.1’s successor could therefore be LTE-M. Having migrated into the 5G standard, its status is secure until the decommissioning of 5G circa 2045. 5G Reduced Capability (RedCap) is also on the way but will not replace LPWA and instead will supplement it, acting as an LTE Cat.1-4 replacement for affordable IoT gateways, routers, and point of aggregation. But RedCap devices are still several years away at the soonest, and there is the complication that 3GPP Release 18 is slated to standardize low-cost RedCap. This seems destined to cause confusion and hesitation on the part of IoT device Original Equipment Manufacturers (OEMs) over what RedCap to adopt and when. All the while, LTE-M will have become the de-facto IoT standard as universally available as LTE is today. Why LTE-M and not NB-IoT? LTE-M already functions in precisely the same way as device OEMs are used to working with LTE, and in the same way as wireless carriers are used to billing and conducting roaming reconciliation. NB-IoT was always designed for stationary devices that infrequently and predictably need to communicate. LTE-M seems the natural candidate for the widest utility. But recent developments may suggest a mixed outcome.

For a long time UNISOC had the only dedicated Cat.1 Bis chipset, the IVY8910DM launched in mid-2019. ASR Microelectronics also offered the less well-publicized ASR1603, which was a Cat.1/Cat.1 Bis chipset with 2G fallback and LTE-M support, having the option to be used with two antennas or with one. It was purportedly the case that, despite the regulatory requirement for a two-antenna configuration, many device manufacturers were already using Cat.1 modules with a single antenna to cut costs. And that the formal allowance of this in the Cat.1 Bis specification in Release 14 was a post-hoc rationalization to legitimize what was already common practice. In December 2021, Sequans became the first non-Chinese semiconductor manufacturer to launch a Cat.1 Bis compliant chip, the Calliope 2 GC02S1. So, perhaps a western brand chipset will assuage western module manufacturer concerns about the technology? Or perhaps Chinese brand Cat.1 Bis chipsets will be fine as long they are wrapped in a trusted western module brand? Thales certainly seems to think so with the launch of its Cinterion ELS62 in January 2022. But what about the limited window of opportunity for Cat.1 Bis? This may not be a problem if a Cat.1 bis module is able to function on 5G Core Network – which is precisely what certain semiconductor manufacturers are now working on.