• I have long held the view that under certain circumstances, today’s AI can improve performance to a superhuman level, and I think that the proposition of Aira Technologies, (see here) a small start-up, is one of them.
• Artificial Intelligence is often heralded as the greatest thing since sliced bread and in some instances, it is, but a lot of the time it massively overpromises and underdelivers.
• Nowhere is this planer to see than in autonomous driving where we were promised self-driving vehicles powered by AI in 2019 but still, they are nowhere to be seen.
• I do not consider vehicles with remote safety drivers to be self-driving as there is still a human element, but it does incrementally move the proposition closer towards true autonomy.
• I have long defined true autonomy where the passenger can be asleep in the back seat of a car with no human monitoring and have no fear for his life.
• The problem can be boiled down to one fundamental flaw which is that AI is not an artificial brain but a sophisticated statistical pattern matching system that is able to find correlations in data patterns.
• In practice this means that an AI system has no causal understanding of what it is doing meaning that it has no ability to adapt should things change.
• Put simply, this means that AI will do a great job in any task where the dataset that defines it is both finite and stable.
• This means that it is feasible to show the AI almost all permutations of possible events and that the dataset will never change.
• This is where using AI for autonomous driving hits a brick wall because the dataset of the road is both almost infinite as well as constantly changing.
• However, this also means that tasks where the dataset is both finite and stable, then a deep learning AI system can perform better than humans.
• Aira Technologies is a good example of this because it uses deep learning to optimise radio packet transmission between a wireless terminal and the base station.
• While the radio environment constantly changes as the terminal moves, the laws of physics for radio transmission remain constant making this an ideal use case for AI.
• Typically, when a packet fails to arrive at the terminal, the base station retransmits the whole packet and will keep doing so until it arrives or the software in the base station tells it to give up.
• Aria Technologies uses deep learning in the digital part of the radio to optimise the retransmission of packets and subpackets such that a substantial improvement in signal range and throughput is achieved.
• Aria has started with Bluetooth as it is open and easy to work with and it has demonstrated a 6db improvement with a bookended solution.
• Bookended means that its technology is in both the terminal and the base station and 6db represents roughly a 4x improvement in the signal.
• This is demonstrated in Bluetooth by achieving a much greater range before the sound quality degrades or disappears.
• This improvement in signal to noise can also be used to provide better battery life at the same range or greater throughput for the same energy budget depending on what the device maker desires.
• Starting with Bluetooth should enable the company to win some customers and generate some cash before considering the bigger opportunities of WiFi or even cellular.
• I have long believed that there are many such examples of where the advantages of AI can be used to their fullest extent without the very limiting drawbacks that deep learning presents.
• This is where I think that the real money in AI is going to be made over the next few years because other areas such as autonomous driving or digital personal assistants are badly suited given the nature of their data sets and hence are likely to continue to struggle.
• This is why outside of these specific areas, I think that AI will fail to deliver on its promises leading to lower investment and falling valuations.
• The hype and chatter on AI have gone pretty quiet since the pandemic and outside of some specific use cases such as this, I expect it to remain that way.

• Qualcomm reported excellent results as the combination of Apple and China is driving a rapid roll-out of 5G enabled devices which is a technology where Qualcomm (as usual with a new radio technology) has the commanding position.
• FQ4 2020 revenues / EPS were $6.5bn / $2.58 nicely ahead of consensus at $5.9bn / $2.23 and above the top end of the rage which Qualcomm had guided for 3 months ago.
• Guidance was also very strong with revenues / Adj-EPS of $7.8bn – $8.6bn / $1.95 – $2.15 well ahead of consensus at $7.1bn / $1.65.
• Not surprisingly the shares rallied around 12% in after-hours trading.
• The main driver without any shadow of a doubt is the wave of 5G roll-out which is progressing more strongly than I, and many others, have expected.
• There are two main drivers are:
  • First, Apple: The iPhone 12 and its variants are all 5G enabled.
  • This means that regardless of whether the consumer wants 5G or not if he or she wants a new iPhone they will get 5G anyway.
  • The main draw of the iPhone 12 is in the usual areas such as the screen, processing power, camera and photographic performance as well as its new design.
  • Hence, while it is these factors that will compel users to make the upgrade, 5G and Qualcomm get to piggy-back on that demand.
  • This is also a very positive sign for 5G in general because as the 5G coverage becomes more widely available, the market will have already been seeded with phones that can use it.
  • Hence, the transition to using 5G will be almost invisible to the consumers and the scene will be set for new services that can make the most of the lower latency and much faster download speeds.
  • What these services are remains unclear but in 4 generations of mobile radio technology, the services have always appeared to fill the space provided.
  • Second, China: China is very keen to lead the race to 5G and as such, a high proportion of smartphones that have shipped this year are 5G-enabled.
  • Qualcomm is pretty well-positioned with the smartphone vendors in China and with the probability that that Huawei starts to struggle in China, this position could strengthen further.
• Hence, the appeal of the new iPhone and the push to 5G in China is driving 5G harder than many expected, and Qualcomm is well-positioned to ride that wave.
• The other area that helps is the RF-front end.
• This is the part of the phone that is responsible for receiving and transmitting radio signals to and from the smartphone as well as converting from analogue to digital and back again.
• With the number of frequencies and standards that need to be supported increasing, this part of the system is becoming increasingly difficult to get right.
• Qualcomm has put a lot of effort here in the fiscal year gone, $2.4bn in revenues came from selling the components and software to manage this piece of the smartphone.
• I think that technical excellence in this area is increasingly important and as others catch up on the 5G radio chip itself, this could help Qualcomm defend market share and show superior radio performance in challenging conditions.
• The outlook for Qualcomm is pretty good as its operations have migrated well to working at home and its end markets are seeing good demand.
• The valuation is high but not in nosebleed territory making this one to hang onto if one already has it.

• The “relaxation” of the US’s stance against Huawei can be taken either as a sign of fear of missing out or as a move to ensure that US companies continue to have a big role to play in the setting of future standards.
• One thing I am almost certain of is that this “relaxation” is not is a prelude to the lifting of any sanctions against Huawei meaning that the rally in Huawei suppliers is not justified.
• The US Commerce Department has clarified its stance on Huawei stating that US companies can collaborate with Huawei on the setting of the standard for future iterations of 5G and other emerging standards.
• While Huawei was not heavily involved in the setting of older standards such as 4G and 3G, RFM research indicates that it has a very strong presence in the 5G working groups which is likely to continue into other technologies and standards.
• If US companies are not permitted to engage in these standards bodies then there is a high probability that it will be Chinese IP that is used even if the US has a superior offering to make.
• Instead, the US is taking the same approach that the Chinese did following the failure of their home-grown TD-SCDMA 3G standard which is to increase its influence in the global standard
• Most people are in agreement that the fragmentation of a communication standard of this nature is bad news for everybody.
• Consequently, I don’t agree with the generally held perception that this is an indication of weakness or a sign of imminent relaxation of sanctions on Huawei.
• This is similar to the position opined by Light Reading (see here)
• Instead, this is a move to ensure that Huawei does not dominate the global standards for future technology correcting an unintended consequence of the sanctions which excluded US companies from standards bodies where they have to cooperate with Huawei.
• That being said, I still think that Huawei’s contributions to the 5G standard have been widely overstated.
• This is because everyone is relying on the number of patents that are declared to have been essential to the 5G standard (where Huawei comes out way in front) as an accurate representation of the real contribution to a standard.
• Patent counting is so fundamentally flawed (which I have argued here) that I do not think that it provides an accurate representation of reality.
• Furthermore, other studies (see here) have found even lower levels of patents that have been declared to essential to actually be essential, casting further doubt over the patent counting method.
• Another method that has been used in the past is based on counting non-self citations of patent portfolios as a way of estimating their importance.
• This method essentially measures how many times other patent applications have referred to a portfolio as prior-art which I have long believed gives a better view of where the real essential patents in a standard lie.
• This method is not without flaws which I discussed back in 2007 (see here), but on balance I still think this is a better method than a straight patent count.
• Unfortunately, I have not seen a non-self citation study on either 4G or 5G meaning that there is no hard evidence to support or undermine Huawei’s claims to the 5G IP crown.
• The net result is that I don’t think that this move by the US Commerce Department changes very much and certainly does not help Huawei’s position.
• This position remains dependent on the much larger struggle between the USA and China where Huawei is an unwitting pawn.
• If there is a reconciliation between the USA and China at a high level resulting in the promised (but increasingly unlikely) phase II trade deal then much of this would go away.
• This is an increasingly unlikely outcome meaning that Huawei may quickly become a Chinese company selling Chinese products for Chinese users in China only.
• Nokia and Ericsson are in the front line to benefit.

• The notion that the number of patents that one holds is a refection of one’s contribution to a technology is deeply flawed and in the case of Huawei, most of the evidence points to the contrary.
• Hence, I don’t think that having a large patent count is going to help alleviate the serious problems that Huawei currently faces.
• A 2019 GreyB study assessed the 5G patent pool (see here) and found that Huawei holds the most 5G-related patents (13,474), ahead of Qualcomm (12,719), Samsung (9,299), Ericsson (8,116), LG (7,694) and Nokia (5,554).
• The researchers then assessed which patents they deemed to be essential and concluded that of the 12,002 patent families that were declared to be essential to the 5G standard, only 1,658 met their criteria.
• Of these 1,658, Huawei holds 19%, Samsung 15%, LG 14%, Nokia 12%, Qualcomm 12% and Ericsson 9%.
• This is being taken as an indication that Huawei, Samsung and LG control the standard and that the older players who controlled the IP for generations 1 to 4 are dropping away in terms of their relevance.
• The key here is how the researchers determined what is essential and what is not.
• GreyB’s stated methodology for this is as follows:
  • “We determined essentiality for each patent family as a Core SEP or not by checking any specifications declared to be relevant by the patent holder to the SEP and compared the specific sections of these to compare overlap of the patent claims with those sections. If partial or no overlap was found, we then broadened our comparison to the wider group of all other specifications to repeat this process”
• My best estimate of what this means is that GreyB compared the claims against the 5G standard and where there was overlap between the claim and the standard, this was deemed to be essential.
• The problem with this analysis is what I have long called “the seats and drinks trolley problem”.
• This issue also came to light in 2007 and 2008 when Nokia and Qualcomm were fighting tooth and nail over 3G IP and patent counting was used to de-emphasise Qualcomm’s contribution to the standard.
• Patent counting is fundamentally flawed and the best way to understand this flaw is to use the analogy of an aeroplane.
• In order for a passenger aeroplane to fulfil its function (transporting people) it requires engines, ailerons and wings but also seats and drinks trolleys.
• Hence the IP for all of these items will be deemed essential to the standard.
• However, the IP for the engines, ailerons and wings is much more essential than the seats and the drinks trolleys because the aircraft is still able to fly without seats and drinks trolleys.
• It is this distinction that I think that this analysis has missed and why I suspect that the Asian companies are over-represented in the SEP count.
• This is for two reasons:
  • First, similarities to 4G: 5G is actually very similar in many ways to 4G and was only created as a separate standard in order to reduce latency and optimise transmission at millimetre wave frequencies.
  • For example, 5G uses the same OFDM coding system that 4G does meaning that the overlap is far greater with 4G than 4G had with 3G and so on.
  • Hence, a lot of the core IP that allows the 5G standard to “fly” is very similar to 4G and may even be replicated.
  • In the 4G patent count, Huawei and Samsung, in particular, have a much smaller share raising the question of how they came from so far behind in a similar standard to lead 5G.
  • One answer could be that they have developed IP that looks more like seats and drinks trolleys rather than engines and consequently, their contribution is less valuable than it appears.
  • Second, Apple: in the most recent patent war, Qualcomm and Apple were engaged in by far the most bitter dispute until Apple suddenly caved in and agreed to pay Qualcomm’s fees and use its 5G chips (see here).
  • I am pretty certain that Apple would have sourced its 5G chips from anywhere else if it would have allowed it to have a working 5G solution for the millimetre wave spectrum that is being used by a number of US carriers.
  • This would have allowed it to carry on fighting Qualcomm and still have a working 5G product for the USA.
  • The fact that it felt it had to use Qualcomm is a strong indication that when it comes to working product, Qualcomm (and I suspect Nokia and Ericsson) has a much stronger position in 5G than this study gives them credit for.
• Hence, I am far from convinced that this study is an accurate reflection of reality.
• Non-self citations (number of times a patent is cited by others) is an alternate method to measure a patent’s value but I have not seen a study of this done for 5G and this method is also beset with issues albeit less than straight patent counting.
• Hence, I don’t think patent numbers will help Huawei much when it comes to its struggles with US (and increasingly the West) as it will have to prove that its IP really is essential which is easier said than done.
• 2020 and 2021 look like they will be extremely difficult years for Huawei as revenues both at home and overseas are at great risk.
• However, if there is a reconciliation between the USA and China at a high level resulting in the promised (but increasingly unlikely) phase II trade deal then much of this would go away.
• This is an increasingly unlikely outcome meaning that Huawei may quickly become a Chinese company selling Chinese products for Chinese users in China only.
• Nokia and Ericsson are in the front line to benefit.

• Qualcomm reported good results as the roll-out of 5G is gathering pace but the uncertainty in China is likely to impact Qualcomm more than most.
• FQ1 20 revenues / EPS were $5.08bn / $0.80 compared to forecasts of $4.83bn / $0.55 but like Apple, it has been forced to widen its guidance range at the lower end.
• Overall guidance was in line with expectations, but the range is slightly wider than usual as the impact of the Coronavirus on both 5G demand in China and the overall handset supply chain remains very uncertain.
• Most factories and workplaces are gearing up to return to work on the 10^th of February, but as the number of cases and fatalities continues to climb, there is some uncertainty around this.
• The central government remains in full control and has threatened severe punishment for any regions that break ranks and allow workers back to work before the centralised say so (see here).
• This could easily drag on past February 10^th.
• While competitors have come up with working 5G modems much more quickly than they did in previous generations, Qualcomm still has an advantage in two key areas.
• These are in the radio front end (electronics which sit between the baseband and the antenna in a handset) and in millimetre wave (28Ghz and up).
• Others have caught up in the 5G modems partially because 5G is actually not that different from 4G, but where most of the changes have been made, Qualcomm still has a lead.
• It remains the only game in town when it comes to millimetre wave which has real issues with signal propagation due to the frequencies at which it operates.
• These are good advantages that will help Qualcomm hold share during the initial rollout of 5G, but it will have to work to keep this edge as 5G technology starts to mature.
• Qualcomm has had a very good run in the last 12 months and compared to other technology stocks its valuation is not yet massively stretched.
• Hence, I am inclined to keep my positive position on the share price for now.

Spotify Q4 19 – Hidden podcasts.

• Spotify reported weak results and doubled down on podcasting as a way to push back against the increasing pressure from Apple Music, Amazon Music and so on, its much larger and stronger rivals.
• Q4 19 revenues / EPS were €1.86bn / LOSS€1.14 compared to estimates of €1.83bn / LOSS€0.77 despite faring better than expected when it came to subscribers.
• Spotify now has 124m premium subscribers, 147m free subscribers (271m total) which it expects to continue growing into Q1 20.
• Part of the problem is that in order to keep growing subscribers, Spotify has to push into lower and lower ASP markets as well as run promotions.
• This has had the effect of lowering ARPU but at the same time expenses remain unchanged which has hurt profitability.
• Spotify also paid the price for being domiciled in Sweden which levies a tax on stock-based compensation when the share price rises meaning that OPEX came in €20m ahead of expectations.
• Spotify also announced the acquisition of Bill Simmons’ The Ringer to improve its coverage of sports podcasting in a bid to create differentiation from Apple Music and Amazon Music.
• The spoken word is an area of growth in terms of time spent and Spotify is right to pursue this vertical.
• However, its user experience needs to be adapted to include this vertical as discovery of podcasts is cumbersome and difficult on its app which has been designed primarily for music.
• Furthermore, its algorithms are designed to categorize and understand music not the spoken word and so it will be very difficult to migrate this functionality across.
• Consequently, I think podcasts will take time to have a positive impact on Spotify’s brand and its profitability.
• I still think that its best shot is to replace the labels which it is quite capable of doing and in that way, there would be a lot more revenue both for artists and Spotify shareholders.
• This is the reason to own Spotify shares and until I see more signs of this I will be staying on the sidelines.

• Making all connected devices compatible is an excellent idea but this is going to be very difficult to achieve as these alliances very often result in a poorly performing standard or abject failure.
• Effectively almost everyone in the connected home space has signed up to form Project Connected Home over IP (PCH over IP) that will develop and promote a new single, royalty-free connectivity standard for devices to communicate.
• The list is long and includes Apple, Google, Amazon, Zigbee Alliance, Samsung SmartThings and IKEA to name but a few.
• Including Zigbee is an excellent idea because many of the traditional domestic electric fixture manufacturers and chip makers are members ensuring that almost everyone in the industry is part of this initiative.
• The project envisages a utopia where the consumer can buy any smart home product from any manufacturer and it will just work and work optimally.
• This sounds excellent but as always with these things the devil is going to be in the details and the different approach being taken many of these companies to the ecosystem is likely to mean a lot of infighting and compromise resulting in a fudged standard.
• For example, Apple monetises the ecosystem via premium device pricing meaning that it can afford to prioritise user privacy whereas Google monetises the data generated to generate revenue.
• These two approaches are fundamentally incompatible, and this is likely to cause problems in the working groups when the standards are being set for security and privacy.
• The project is going to be overseen by the Zigbee Alliance which should ensure impartial and objective management, but there are still some large pitfalls to avoid:
  • First Glacial slowness: Projects like these set their standards in committees.
  • Typically, the way it works is that the participating representatives from the members are split into a series of committees each of which deals with one aspect of the standard.
  • These are then put together at the end to produce a single working standard upon which everyone has agreed.
  • This is a very slow and arduous process and unless this is expedited by Zigbee, will take a very long time to produce anything.
  • Second, Performance: which is usually the biggest problem.
  • The members of each committee have different agendas which are to ensure the best possible outcome for their company’s products.
  • This means that each member will have a preferred implementation that suits his product better and very often this will differ between the members.
  • What often results in these situations is a fudged compromise where the technology has not been implemented to suit the preferences of the committee members rather than optimal performance.
  • This leads to suboptimal performance from the standard which has often led to its failure.
  • The many attempts to create Linux and Java software for mobile phones between 2000 and 2007 are great examples of this.
• If everyone in the industry starts to use this standard then its weaker performance may be less of an issue, but there are a number of players: Qualcomm, Broadcom, Xiaomi and so on who have a lot of involvement in the smart home and are not present here.
• These could come up with proprietary systems that outperform the open standard causing manufacturers and consumers to think twice about its adoption.
• Hence, PCH over IP needs to avoid compromise and fudge wherever it can and the responsibility for this lands squarely on the shoulders of the Zigbee Alliance.
• I am sure the Zigbee Alliance is fully aware of these issues but whether it has the bench strength to ensure a good outcome remains to be seen.
• If this works, it will be great for the consumer, but history leads me to be cautiously optimistic at best.

• Verizon’s 5G coverage map in the USA (see here) shows that the only time users are going to get reception is when they are outdoors with line of sight to the basetstation and, most likely, stationary.
• Verizon’s 5G coverage map in the 18 US cities that it covers reveals extremely patchy coverage where it appears that each basestation is only capable of providing service within a short distance with line of sight and with no indoor coverage.
• Furthermore, it also appears that the presence of trees also blocks the signal as evidenced by the coverage of Battery park (see above).
• In this example, it would appear that there is a basestation on State Street that extends some coverage into the park itself.
• When one looks at a satellite image of Battery Park it becomes clear why only a portion of the park is covered as there is an empty space next to State St (see here) and then a line of trees.
• It looks like it is the trees that are preventing the signal from propagating further into the park.
• This comes as no big surprise because Verizon is deploying 5G at 28Ghz and 39Ghz which are in the millimetre wave band.
• This refers to frequencies between 28Ghz and 300Ghz which is where 5G has really been designed to operate.
• The key differences between 4G and 5G that make them separate standards are coding changes that have been implemented to ensure very low latency and high efficiency in millimetre wave frequencies.
• Millimetre-wave is great for throughput and high capacity as the extremely rapid oscillation of the radio wave means that vast amounts of data can be transported at very high speed.
• However, the trade-off is that propagation (how far the radio wave can travel before its energy is absorbed by the air or obstacles) is very poor.
• This is why large numbers of basestations are required which can only offer very limited coverage outdoors and almost none at all indoors.
• This probably means that there will also be no in-vehicle coverage either.
• Consequently, Verizon’s 5G service is going to be next to useless because, in almost all use cases where a user might want to make use of this bandwidth, there will be no coverage.
• This is why most operators are also rolling out 5G at frequencies between 6Ghz and 3Ghz.
• This will give much better coverage, but one also finds 4G at these frequencies which will provide comparable performance given how similar the standards are.
• Hence, the added value that 5G brings to the table for the user is highly questionable.
• I continue to think that the real impetus for rolling out 5G is fear of being left behind and the reputational loss that might occur as a result.
• This is being exacerbated by the growing rivalry between the USA and China where China is rolling out 5G as fast as it can.
• This will put pressure on US operators to accelerate their roll-out in spite of the low or negative return that will result.
• The real winners here are Ericsson and Huawei who are winning the lion’s share of this spending.
• Limitations on Huawei’s infrastructure business look likely to be limited to the core network rather than the radio access network and it is in the radio where most of the spending is to be found.
• Ericsson is where I would put my investment exposure in 5G if I was forced to invest in this theme.

Patent counting is very misleading.

• A report at the RISE conference in Hong Kong lays claim to China’s superiority in 5G due to the fact that it has more 5G related patents than anyone else.
• However, reality tells a different story which underlines the perils of using patent counting as a measure of value.
• A report by the South China Morning Post and Abacus (see here) highlights the flow of innovation in apps from China to the west but also lays claim to superiority in 5G.
• This is based on a patent count where China has 3,400 5G related patents, South Korea 2,051 and US 1,368.
• This suggests that China is the leader in 5G with South Korea in second place.
• This is at odds with reality where the only working commercial 5G handset radio chips are from Qualcomm and Samsung which one will find inside every Chinese and South Korean device.
• If patent counting was a reliable measure, then there should have been plenty of Chinese and South Korean 5G radio chips but there are not.
• The reason for this is that not all patents are created equal where some are worth hundreds of millions and many are worth nothing.
• In order to understand this, one can broadly classify patents into two categories.
  • First, Standard Essential Patents (SEPs): These are patents that have to be used in order to be compliant with a standard such as 5G.
  • As a result, it is much easier to prove infringement for SEPs but at the same time, the holder is bound by the principle of Fair, Reasonable and Non-Discriminatory (FRAND) licensing.
  • This basically means that one has to licence the patent to anyone that wants it and must do so fairly and not take advantage of the fact that companies building products have to infringe the IP to implement the standard.
  • This is what Qualcomm’s current case with the FTC is all about.
  • Second, Implementation patents: These are patents which are not essential to any particular standard and stand on their own merits in terms of the value they deliver.
  • They are much harder to prove infringement for but critically there is no requirement to licence nor is there a limit on the licensing fee that can be charged to use it.
  • Furthermore, injunctions may be sought as a remedy for infringement.
• The best way to understand how patents differ in value is to use the example of a house.
• A SEP would say that the house has to have windows.
• The implementation patent would then determine what the windows are made of, what shape they are, whether they open and so on.
• The patent that says windows are made of glass is extremely valuable but the patent that says the windows are made of pressed steel clearly is not.
• This is why merely counting the patents that each country has filed related to 5G is meaningless.
• The beauty of patent counting is that its easy to understand and easy to carry out.
• The problem is that it has led an arms race whereas many patents as possible are filed as possible meaning that finding the good quality IP becomes like looking for a needle in a haystack.
• Qualcomm pulled out all the stops to get 5G working (like it did for 3G and 4G) and as a result, has got to market first and therefore may have created a good portion of the really valuable implementation patents.
• This was also instrumental in its defeat of Apple in their long-running licencing battle (see here)
• Consequently, I don’t think that the real technical expertise in 5G is to be found in Asia and that this patent count provides an unrealistic picture of the landscape in 5G.
• I still think that the take-off of 5G is going to take longer than many expect due to the lack of real use cases that distinguish it from 4G (see here).
• However, in that light, Qualcomm is the one I would look at for an investment in 5G if I was going to make any at all.

• Verizon has been forced to ditch the $10 premium that it charges customers to access the 5G network as the performance delivered is exactly what the laws of physics predicted.
• Verizon, of course, has dressed this up as a sales promotion but this fools no one.
• The reality is that frequencies of 28Ghz and above (where Verizon is operating) offer incredible capacity and speed (as observed) but very poor signal propagation.
• This is because the energy in the radio wave is being consumed very quickly because it vibrates so rapidly, meaning that the entire signal is often absorbed as soon as it hits anything.
• It also means that the radio wave will not travel very far from the base station (as observed) before its energy is completely depleted.
• The net result is that when one is stationary within line of sight of the base station fantastic performance is delivered but the minute one moves or even turns one’s back to the transmitter, things go rapidly south.
• This is why early users have reported blazing fast internet when they can get a signal near a base station but as soon as they move the signal is quickly lost.
• Furthermore, tests also showed that there was no sign of the supposed 1ms of latency with latencies in the range of 20-30ms being observed (see here).
• This is, of course, a very limited data set, but the promise of 5G is to deliver consistently superior latency which in this case it is not.
• Early performance of 5G is clearly inferior to 4G on speed, latency and jitter as BTIG and RFM’s indoor tests show (see above).
• Verizon now has a nasty chicken and egg situation.
• It badly needs to roll out thousands and thousands of base stations to deepen its 5G coverage so that it becomes worthy of the $10 premium, but it will be reluctant to spend all this money with no visibility on users’ willingness to pay.
• Its peers are in an easier position as they are operating their initial roll-outs at the much more forgiving frequencies below 6Ghz, but it is still a mystery to me how this service will be better than gigabit LTE.
• The net result is that early signs are that 5G is not nearly good enough to command a premium price which combined with the early stage of the terminals (see here) means that volume is going to be tiny.
• I suspect that this leads us straight back to what has always been my core use case for 5G which is the very boring but economically exciting fixed wireless access in the USA (see here).
• Hence, I think that the real value of 5G to an operator at this stage remains bragging rights and nothing more.
• This will result in an initial roll-out of thin coverage and then a pause while everyone searches for a use case over 4G and the while technology matures.
• Once this is in place an orderly roll-out will follow and over time, 5G will eventually replace the much older 3G and 4G networks.
• However, much hype and promise have been made on the instant success of this technology and so there will be a period of disappointment and falling valuations as the laws of physics reassert their supremacy.

• MWC is almost over and while the stages have seen some confident predictions (eg Cisco: “hype is going to be reflected in the reality”), the message from the floor and in the trenches is much more realistic.
• To me, this is excellent news because, in reality, no one is getting over excited and consequently the bar which 5G has to get over during its early stages to meet expectations is much lower as a result.
• This means that there will be a much smaller correction when the pause comes compared to the situation 20 years ago when 3G was going to be amazing out of the box.
• The reality is that 5G is early stage and it has significant issues.
• However, these problems are the same problems that we had with both 3G and 4G which leads me to expect that they will be solved more quickly than before:
• These include:
  • First, bulky handset: this is largely caused by the lack of integration of the core 5G components and a bigger battery (see below).
  • While the kinks are still being worked out of a system, one keeps the components discrete so that changes can easily be affected.
  • Once everything is working well, then the integration phase begins and the devices become smaller and lighter as the 5G footprint shrinks.
  • This will begin in 2020 with Qualcomm most likely to be the first to launch integrated 5G products.
  • Second, heat and battery life: RFM tests of live 5G products quickly showed that both basestations (where they could be reached) and the devices were running very warm.
  • This is largely due to the power amplifiers which have once again fallen back to around 15% efficient with the new technology.
  • This means that 85% of the energy goes in is dissipated as heat, killing the battery life and making the devices run hot.
• These problems are nothing to get worried about as they are the exact same problems that plagued the early stages of both 3G and 4G.
• However, it does mean that the 1^st iteration of devices will be bigger, more expensive with lower battery life than their 4G only counterparts.
• The problem with this is that from a consumer perspective, 4G is already so good that the user may not notice any impact from 5G in his Digital Life.
• Consequently, other than bragging rights, there is little reason for a consumer to pay up for a 5G device in the short-term.
• The net result is that 5G is pretty much where I thought it would be and I continue to expect an initial roll-out followed a pause while the technology matures and everyone works out what to do with it.
• Once the base has been re-established, I would expect 5G to gradually take-over other radio technologies leading to slow but steady growth in infrastructure and handset or terminal sales.
• The good news is that because deep down, the industry knows this, meaning that the pause is already anticipated and expected.
• Party animals will be feeling a bit miffed at this MWC but no party also means no hangover.