The 5G business case is a hot topic, especially with emerging markets operators. This week Strategy Analytics said that device subsidies will be necessary to build penetration. No one should be entirely surprised.
I saw the report in the GSMA’s daily newsletter. It quotes the analyst firm as saying that 5G smartphones will be the most complex and expensive mobile devices ever with a wholesale price of US$750 (yes, the decimal is in the right position; yes, that is the wholesale price). The article speculates about retail pricing but that really is the key question: can these be sold as retail devices at any margin or must they be sold only at cost or, more likely a discount?
A wholesale price of that order of magnitude would put a 5G device in iPhone X territory, without necessarily having ‘hero-phone’ attributes to go with the ‘hero-phone’ pricing. How much would a 5G iPhone X cost?
Worse, Strategy Analytics believes that the normal price decline with time – the experience curve – would be slower than that of 3G or 4G. I suppose if the curve starts with a price very much higher than 3G or 4G it is going to take longer to come down to levels that are competitive with current devices.
The GSMA quotes Ken Hyers, SA’s director of emerging device technologies saying, “It requires magical thinking to expect that consumers are going to rush to buy 5G smartphones that are bigger, and more expensive than any phone that they’ve ever bought before.”
Now before getting into the implications for 5G, especially in emerging markets, I have to point out that this is the complete opposite of what I heard at 5G America’s Analyst Forum a few weeks ago. I know Ken but do not remember seeing him there, so he may have formed his own opinions without the benefit of what we were told (or maybe despite what we were told).
At that event, Sprint said that a 5G smartphone below 6Ghz was no more complex than an LTE Advanced phone. Hero phones with LTE Advanced sell for around US$1,000 but there are cheaper models available.
Device manufacturers said they expected the experience curve to be faster (not slower) than 4G.
Another caveat to Ken’s comments may be that he refers to a so-called mm-wave device – one that operates at wavelengths less than 1 mm or equivalently at a frequency well above the 6Ghz talked about at 5G Americas. The ranges normally being discussed are around 30Ghz. The GSMA article does not mention anything about frequency and neither does the original press release. The distinction is important.
Such devices do not exist yet except in ‘breadboard’ form and, so far, are thought to have significant challenges with the necessary MIMO. We have solved the issue of having 4 antennas in a ‘normal’ smartphone without it looking like a hand-held porcupine, but 16 or 32 antennas are literally orders of magnitude much more complex problems. Ken’s commentary about the phones not only being more expensive but physically bigger suggests he is thinking of this problem.
As (I hope) is well understood, 5G is not a monolithic upgrade as previous generations have been but a series of technologies that address a number of canonical use cases. The canonical use cases have been grouped into three axes:
Skipping over the details, I hope it is clear that these have different characteristics and while there maybe common elements to their solution, there will likewise be solution elements that are used uniquely for only one or only two of the three cases.
Thus, many of these technologies overlap but, frankly, others do not or at least there are technologies that are used only in certain situations when the use case demands it. For example, higher-order MIMO is needed for mm-wave implementations and mm-wave is needed for use cases that require very high bandwidth.
Some capability may only be needed much later. Again, mm-wave for mobile use cases is not perceived to be required in the short or even intermediate term. Just like 4G was an enhancement to 3G in its first generation and not something dramatically different, 5G eMBB below 6Ghz will be ‘enough’ for a while.
Which is why it matters if Ken’s comments are referring to sub-6Ghz smartphones or mm-wave smartphones (assuming such a thing is even possible or needed).
Let’s assume I am splitting hairs way too finely. Let’s assume that Strategy Analytics believes that the smartphones needed to enable the eMBB use case will be “most complex and expensive mobile devices ever with a wholesale price of US$750”. The operators likely to experience this challenge are AT&T, Sprint and T-Mobile (late this year and into next) and then Korean and Japanese operators.
What would the consequences be?
I think it fairly obvious that if everyone needed a ‘hero phone’ and had to pay ‘hero phone’ prices to get one, then demand would be severely restricted. There are certainly a lot of iPhones in rich markets like the US (and Galaxy S9s and etc.) but there are also a lot of down-market devices amongst those iPhones. Not all are iPhone X’s by a long stripe.
Hence SA’s assertion that we are back to subsidized phones.
CFOs and shareholders will not be pleased. Nor, in some markets, will regulators who view subsidized phones as anticompetitive and an inappropriate cross-subsidy. The retail channel gets complicated when there are subsidies to be administered, not impossibly so, but channels that formerly did not have to deal with an operator at all will be unhappy.
And implicit in a subsidized phone is a way to recover the subsidy, i.e. an expectation that those who receive the subsidy will have a higher ARPU than those who do not receive one. Which somewhat suggests we are back to another example of potential “magical thinking” as Ken says: that people will pay more for eMBB on 5G.
There are, of course, two ways to pay more: pay more for the same and buy more. The former has never worked: few have been willing to pay more for the same package just because the ‘G’ changed. But we have successfully convinced people to buy more, adding data packages to voice-only subscriptions or incrementing their GB’s because their lifestyle requires more data and they do not want to run out. More bandwidth has meant more applications, more consumption – especially more video – and, when price is managed correctly, higher ARPUs.
Or sometimes just ARPUs that do not fall as fast as they did before.
This all gets worse, of course, when we are talking about emerging markets. Apple’s share for new top-of-the-line models is in the low single-digits. There are a lot of refurbs, hand-me-downs and iPhones of questionable provenance circulating but it will be a while before the 5G grey market gets to the point where it can have a noticeable impact on adoption.
If 5G phones are going to be the price of a new iPhone X then either the operators are going to have to heavily subsidize or 5G eMBB is not going to happen. Maybe nowhere but certainly not in emerging markets.
Which is why I have been conservative about my expectations for 5G in the region and bullish on advanced levels of 4G. (And on B2B applications of 5G i.e. the Massive IoT and URLL use cases.)
But if this scenario about device prices is correct, then 5G may not be an opportunity for emerging markets to leap frog.
Instead it may be the generation that gets skipped.
(Title Reference: I have to admit I had almost forgotten this Leo Sayer song that had originally been recorded by Roger Daltry in 1973. Hearing it again, I think I remember the Leo Sayer recording more (and that is barely). The phrase was floating around in my mind as I wrote this and a bit of Dr Google turned up this reference.)
No Comments »
Leave a Reply