The curious case of IoT infrastructure
Pol Ferrer Catasús (Investments)
This week is the MWC in Barcelona, and so, that time of the year when telcos, OEMs, equipment vendors and pioneering start-ups alike praise the spells of IoT. For good reason. With so many interesting use cases to connect anything, from alarms, to freezers, to trees, to cows, it is easy to captivate one’s imagination. In this article, we aim to provide a dose of realism to the hype that surrounds IoT, and that will only accentuate as the MWC days go by. This perspective is drawn from our experience seeking attractive opportunities for our growth equity fund, and attempts to challenge some of the claims made by those who live from connecting our lives. As much as we appreciate the level of sophistication (or lack thereof) of some of the technical solutions that have been devised for the IoT, we cannot but recognise certain flaws in the premises upon which these are being built:
- The snobbery of trying to build the first global network: As a matter of fact, none of the largest mobile operators, despite their hefty CapEx budgets, reasonable profitability levels and longstanding market leadership have ever built a global network. This is partially the result of limited and expensive spectrum and the high costs of network deployment, which LPWA players on unlicensed spectrum claim to have overcome, but also of regulatory pressures and competitive dynamics that IoT infrastructure players also need to abide by. Aiming for global dominance from the onset is, in our view, a wild-goose chase.
- ... While assuming that others will pay for the deployment: When you have nothing and want to conquer the universe, management books suggest that you don’t go it alone. However, it is an illusion to think that one can outsource the costs while only reaping the benefits. Many parties may have an interest in deploying LPWA networks, from tower companies longing to become service providers, to telcos concerned about missing the wave of IoT, and we have indeed seen some notorious pioneers. However, faced with aggressive deployment targets and equipment purchase commitments, these considerate athletes may struggle to monetise their empty networks with a per-device revenue model.
- The overstated need of a new standard: IoT infrastructure providers predicate the need of a new standard that can address the myriad of use cases that compose the Smart Society. In reality, existing technologies go a long way in enabling many of the smart-home, connected-car and connected-self applications that have come to light in recent years. As a matter of fact, most of today’s consumer IoT solutions, from alarm systems to fitness trackers to thermostats, are based on a combination of Wi-Fi and Bluetooth, with cellular back-up when required, and therefore only consume bandwidth from telecom services we are already paying for. While we think that industrial and government applications stand to benefit from more efficient protocols, we are sceptical about their potential mass adoption in the consumer segment.
- The promise of the 50 billion connected devices in 2020: Since Ericsson’s prediction of 50 billion connected devices by 2020 six years ago, technology analysts, strategy consultants and other futurists have produced thousands of pages justifying ostensibly high numbers. Regardless of the ambiguous definitions on what constitutes a “connected device”, the reality is that, as of today, LPWA user equipment is still in the couple of dollars, which directly precludes many of the highly ingenious use cases that are being conceived. While equipment prices are falling fast, it is uncertain they will reach the sub-$1 level, the tipping point for widespread adoption in our opinion. This number, of course, should account for connectivity costs, which are precisely those generating the hefty revenues IoT infrastructure players are forecasting based on utopian pricing models that may require a revisit.
- The fallacy of “We build the network, applications will come”: Although some LPWA vendors have already achieved high double-digit coverage in multiple markets, their networks are markedly underutilised. The technology is available but unfortunately not the demand. We believe this is partly explained by the limited number of off-the-shelf applications that can seamlessly plug into these networks. Infrastructure players do not have the capabilities and, until recently, the interest to build and commercialise IoT applications. Perhaps they should? Christensen’s theory of interdependence and modularity, which indicates that companies should insource product development when existing solutions are not good enough, seems to point in that direction.
- The marketing tag of the number of partner operators: We often hear this recurring comment to express the belief that such LPWA technology is closer to becoming the IoT standard because of the imposing number of telecom operators endorsing it. However, the reality is that operators have just been hedging their bets and partnering with everyone. While expensive, this may not be an unreasonable strategy given the uncertainty around which technologies will prevail. For instance, SK Telecom is an investor in Sigfox, a founding member of the LoRa Alliance, and has launched an LTE-M network. What else?
- Unlicensed spectrum? Yes, we can: Unlicensed LPWA is predicated on the assumption that it will be viable to use ISM bands at scale for commercial purposes. It is also based on the premise that IoT infrastructure players will be good neighbours to other broadcasters on the same bands. This, of course, is a working hypothesis, as no large-scale real-life deployment with thousands of devices per access point has ever been conducted. While we conceive that one can stealthily use citizens’ spectrum in sub-scale applications, we do not take it for granted that one single player will be allowed to saturate an ISM band for its own commercial purposes without anyone complaining. We expect regulators’ oversight to increase as they realise that this was not the purpose these bands were reserved for in the first place. We do not have a technical opinion on whether there will be interference between LPWA users when the number of connected devices scales up, however, we note that even if there is none, the mere suspicion of interference can already cast a shadow on adoption.
- We’ll always have... Data: We have recently heard of a few players that, when challenged about their monetisation strategy, invoked the holy grail of customer data, as if the vast amount of information that could be collected in a connected world was the deus ex machina that could protect any business model when everything else falls apart. Not only are we wary about these Google wannabes, we are particularly concerned with the unaddressed privacy vulnerabilities that a connected world exposes us to. We reiterate our view, expressed in previous articles, that protecting our lives from “Big Brother” presents an unresolved market opportunity. We have yet to see market driven solutions seeking to preserve this fundamental principle.
All in all, we too believe that the oncoming wave of IoT is unstoppable, and we are excited by many of the use cases presented to consumers, businesses and governments to improve decisionmaking and productivity. However, we also believe that the irrational exuberance displayed by winner-take-all panaceas should be grounded with some fundamental questioning. If you are currently taking part in the global IoT race, and are on the verge of signing multi-year contracts or committing top dollar to nationwide network deployments, we suggest you challenge your vendors to confront our humble observations, to avoid getting caught off-guard at a later date, when things don’t quite unfold as expected.
© 2019 Delta Partners.