How To Solve The Cellular IoT Prisoner’s Dilemma

It seems like we are finally getting clarity on features and pricing not just for high powered cellular IoT like LTE CAT M1, but also the long-awaited NB-IoT as well. NB-IoT famously claims to be a low power, wide area networking (LPWAN) technology.

This piece in Lightreading: “NB-IoT Gets Insecurity Complex” builds on other research we have seen that basically confirms that not only are NB-IoT modules not going to price at $1 each (what we were told just six months ago) or even $5 each, but much, much more:

“A likelier explanation is that customers are not ready for NB-IoT prices. Modules still cost somewhere between €10 ($11.25) and €15 ($16.87), according to Deutsche Telekom, against an industry target of just $5.”

It is worth noting that the article quotes pricing for NB-IoT “modules” — not completed devices. And NB-IoT modules are arriving at price points that are 2–3x or more over what was promised. Assuming the lower end of Deutsche Telekom’s per-module cost of $11.25 and adding additional bill of materials, packaging, shipping, taxes and other costs required to ship a completed device, and you likely end up with per endpoint costs of $20, probably more. And that’s before the concept of profit even kicks in. The commercial/industrial appetite at these levels, even amortized via a monthly subscription, is quite limited.

As I’ve mentioned before, my experience with IoT products that come with this kind of premium pricing is that they self-segment into corners of the market where ease of use/ease of implementation cancels out the sins of high device or monthly subscription costs. Certain small business and consumer segments are the most likely targets for a premium product like this, while commercial and industrial customers with rudimentary systems integration capabilities and who don’t necessarily trust their carrier with their data will be slow to adopt. Obviously, the “forking” going on in the NB-IoT standardization community does little to assuage the worries of enterprise customers:

The industry’s reactions to those reports are what betray the insecurity. Talk of interoperability problems between equipment vendors Ericsson AB (Nasdaq: ERIC) and Huawei Technologies Co. Ltd. is now rife at industry events. Numerous industry experts have also flagged the issue in discussions with Light Reading. Yet Ericsson and Huawei have pleaded ignorance, while operators have either denied there are problems or declined to comment.
The timing of one recent statement on interoperability also looks odd. Vodafone this week said it was carrying out NB-IoT interoperability tests more than two months after it was supposed to have launched commercial services in some European markets, and having never previously acknowledged that interoperability is a concern. Those tests have shown that all is tickety-boo, it insists.

But assume that the standards folks get their work done sooner or later. Ease of use or ease of implementation still assumes, in the LPWAN space, that the customers can “plug and play” an IoT device (or thereabouts) and not need to worry about maintenance, etc. Which brings us to the money question: battery life performance. Dead batteries are the maintenance bane of IoT network operators everywhere. For now, NB-IoT is deploying TCP/IP as its networking stack, and TCP/IP is no friend to battery life. A real risk, therefore, for NB-IoT is that it ends up as a high powered backhaul or mains-only solution similar to its LTE Cat M1 cousin which is also showing its true colors as a high priced/high power solution. Or NB-IoT ends up as a repeat of previous high powered, high priced consumer GPRS misfires.

If NB-IoT does not deliver multi-year battery life as promised — and I do not believe it will — the cellular industry has a bit of a situation on its hands.

NB-IoT is really a “we-can-just-do-a-simple-base-station-firmware-upgrade-and-deliver-low-power-wide-area-IoT” pitch to analysts who track LTE. Free money! But without a technology to actually, you know, deliver the broader LPWAN market opportunity, the pitch is moot. So analysts might rethink their cellular LPWAN forecasts absent one or more of the following moves:

  1. Embrace an additional, non-LTE radio option like LoRa or Sigfox. Or another emerging LPWAN PHY layer technology. Even if only as an “interim” solution until the LTE-based LPWAN woes can be resolved, assuming they can. If there’s a market worth going after — and there is — it’s probably not worth sacrificing early market share to wait years for your preferred radio technology to be ready for prime time.
  2. Align per-month endpoint pricing to provide something closer (on a present value basis) to parity for a comparable, easy-to-use technology from an unlicensed band, e.g. LoRa. While this could help preserve earlier unit forecasts, it may be get howls from cellular finance departments. But if you can’t compete on battery life, at least soften the price blow if you are serious about competing for customers.
  3. Re-orient internal mindsets from steadfast allegiance to LTE and instead embrace a more open-ish model, perhaps along the lines of Made for iPhone, where carriers embrace multiple radio technologies and devices and provide a common networking layer to minimize maintenance and support. Like this one.

Others have offered suggestions to help carriers out of this dilemma, and the answers don’t appear simple. And while I shouldn’t be surprised, I am nonetheless astonished that the LTE community has even come to this point given its massive resources and market power. What happened? Here’s an excerpt from the same piece:

What seems undeniable is that NB-IoT was a rushed job following an abrupt rethink by the cellular industry on the need for a so-called low-power, wide-area (or LPWA) technology. Two and a half years ago, cellular industry folk at the GSM Association (GSMA) were “dismissive” of LPWA, according to Tom Rebbeck, a director at the Analysys Mason market research business. “Then within a year they had turned around because they saw the momentum behind Sigfox and LoRa,” he says.

I won’t underestimate their ability to turn the ship around, but based on what I hear from colleagues working on 5G (coming to your town in how many years? Five? Ten?), the cellular industry mindset appears to be one of doubling down on the current roadmap, which will be no friend to low power endpoints.

So maybe low power IoT was not so important to the cellular community after all and all the hype was … just that. Regardless, this represents a great opportunity for non-LTE LPWAN radios, LPWAN networking technologies, and network operators not saddled with LTE-or-die obligations. And it’s still early enough for carriers to re-think their plan.

If you liked this article, please hit the ❤ icon below and follow me.