Google & Amazon – Second blood.

Reply to this post






Second blood goes to Google.

  • Google has scored a badly needed win over Amazon as Amazon has said it will start selling Google’s Chromecast devices once again which is almost certainly a bid to get YouTube back on its devices.
  • This is a sure sign that Amazon needs Google more than Google needs Amazon and further underpins how important the ecosystem is.
  • When I compare Google’s digital ecosystem against Amazon’s the results are night and day.
    • First, Digital Life: On RFM’s Digital Life analysis Google scores 40% while Amazon is on 30%.
    • However, if I look at the services that have hundreds of millions of users, then Google is miles ahead as outside of shopping the only service where Amazon has traction is Media Consumption (Amazon Prime Video).
    • It is clear that Google’s ecosystem is far more developed and more complete than Amazon’s.
    • Second, RFM 8 Laws of Robotics: These are the measures that asses the quality of these services and here the contrast is even more stark.
    • Google has problems around the user experience on Android driven mostly by software fragmentation and an inability to update software but outside of that it gets top scores across the board.
    • Amazon’s score indicates quite clearly that it has no real understanding of what is required to build a thriving ecosystem that users love and demand.
  • The result is that users of Amazon devices are likely to demand that Google services are made available on them with a good user experience and are less concerned with Amazon’s own services outside of shopping.
  • Failure to provide this could easily drive those users to buy other products as Amazon’s one Digital Life service that has some traction (Amazon Prime Video) is widely available on most streaming devices and smart TVs.
  • This is what has given Google the upper hand in this latest spat and I am certain that this is why Amazon has backed down.
  • This is a badly needed win for Google as in the smart home its digital assistant remains way behind Amazon’s Alexa despite being a superior product.
  • I still see Google as being on the backfoot when it comes to the smart home, but it has closed some of the gap to Amazon in terms of third party developer support and it remains a superior product.
  • This will be a key battle that is played out in 2018 and the level of support offered by device and service developers when they show their wares at CES in January will be a key indicator.
  • Market penetration remains very low and so there is still everything to play for but I still think that in the long-run Google should win as it has the better product.

Apple – Super cycle blues.

Reply to this post






Early data indicates that there is no super cycle

  • While it is too early to say for sure, the initial data suggests that the super cycle that the market is looking for to underpin the valuation of the shares is not occurring.
  • Mixpanel is an analytics company that tracks user interactions for the providers of apps and services on the web as well as mobile devices.
  • With 20,000 customers and 7tn data points collected every year, I think it is fair to say that Mixpanel is quite capable of collecting a data sample that is a good representation of reality.
  • The latest data from Mixpanel tracks the appearance of iPhone 8/8+/X devices in the data that it collects, giving an idea of how quickly the user base is switching to the new models.
  • In the two months since release (iPhone X one month only), the three devices have racked up the following share of the user base:
    • iPhone 8: 2.79%
    • iPhone 8+: 3.68%
    • iPhone X: 4.91%
    • This gives total penetration of the new generation of iPhones in the first two months of availability of 11.3% of the user base.
  • In comparison, the iPhone 7 and 7+ penetrated 23.5% of the user base within the first two months of availability.
  • This strongly suggests that the latest generation of devices is not generating the kind of super cycle that the market is looking.
  • I do not think that supply is an issue as I can order an iPhone X 64GB today and pick it up from the Apple Store in New York City or Palo Alto tomorrow. The iPhone X 256GB wait is four days.
  • To be as consistent as possible, it is necessary to adjust the data to reflect the fact that the iPhone X has only been available for a month.
  • If I take the penetration achieved in the first month by each model in the new generation, I end up with a total user base penetration of 8.8% after one month of availability.
  • By contrast after one month the iPhone 7/7+ had achieved 11.5% penetration of the iOS user base.
  • This is a strong indication that uptake by users of the new generation of devices is actually slower than it was for the iPhone 7.
  • It is important to remember that this is going to be offset by price as the data clearly shows that the iPhone X is much more popular than the iPhone 8, despite being 30% more expensive.
  • Hence, softness in unit shipment numbers could easily be compensated by the higher price being paid for the iPhone X.
  • However, I think that expectations for Apple for FY2018 are including both a strong replacement cycle and higher prices which is why I fear disappointment in the short term.
  • The caveats to this analysis are:
    • First, data: I have assumed that Mixpanel is a statistically significant representation of reality.
    • While the data quantity suggests that this should be the case, I have no hard data to prove it.
    • Second, availability: I have tested availability in NYC and California as they are places with the highest demand, but this may not be a true representation of supply.
    • Third, user base: The user base of iOS devices is larger this year than it was last year.
    • Consequently, it will require more devices to be sold in order to achieve that same level of penetration as the iPhone 7.
    • However, I do not think that differences in the size of the user base is enough to account for this discrepancy.
  • When I weigh up the findings of this analysis and take into account the caveats, I conclude the following:
    • First: uptake of the new generation of iPhone is slower than last year and there is no sign of the much hoped for super cycle.
    • I think this is mainly driven by the very high price of the most desirable product which will lead to many users waiting until the full screen penetrates to the lower priced segments.
    • Second: the higher price of the iPhone X will offset the impact on Apple’s revenues of a slower uptake compared to the iPhone 7, but it will not enable the company to soundly beat market expectations.
  • Consequently, I think that while demand for the iPhone X is clearly quite high, it is not high enough to allow Apple to beat the already very high financial expectations that have been set for it.
  • Hence, I can see disappointment coming through in Apple’s next few sets of earnings releases and think that the time has come to take some money off the table.
  • Tencent, Baidu and Microsoft are worthy candidates for consideration.

Faraday Future – First to fall.

Reply to this post






FF looks certain to fail. It won’t be the last.

  • The saga of Faraday Future may soon draw to a close, which combined with the blacklisting of its biggest investor (Jia Yueting) on China’s national debtors’ database, is a strong indication that this group will never produce a vehicle.
  • Faraday Future was one of the first Chinese companies to follow Tesla’s lead into electric cars but it has been quickly followed by others such as Nio and Byton, all of whom have big ambitions to benefit from the shift from petrol to electric vehicles.
  • However, the last two years for Faraday Future has been a litany of woe which I continue to think will result in the complete collapse of the venture.
  • It turns out that building vehicles (even electric ones) is quite difficult and when one is beset on all sides with law suits and angry creditors (see here), one ends up with huge turnover of management.
  • This means that virtually no attention is being paid to the development of a vehicle and certainly no production.
  • Furthermore, I think Faraday Future has very little chance of raising any more money as no investor in his right mind would back a company with the following characteristics:
    • First: Its main backer has been put on a credit blacklist for failing to pay RMB462.1m that he owes to Ping An Securities.
    • Jia Yueting is currently residing in the US and appears unwilling to return to China.
    • Second: Faraday Future’s sister company LeEco is stranded in limbo after having to abandon all of its international ambitions due to over extending itself and getting into financial difficulty.
    • The shares of the parent company Leshi Internet and Information Technology have been suspended in Shenzhen since May 2017.
    • Third: Previous employees report that the company barely has enough money to pay its staff and that financial controls and systems are inadequate for a company of this size.
    • Fourth: Faraday Future’s recent hire Stefan Krause who was brought in to sort the financial situation out left the company after just 6 months having demanded changes that were not to the liking of Jia Yueting.
  • Furthermore, there are now serious doubts as to whether Faraday Future will ever make it to the new and smaller manufacturing facility in Hanford, California.
  • This was the new and smaller location after the much bigger site in Nevada never broke ground as the contractor was not paid.
  • All of these problems come before the difficulty of designing and producing a vehicle which is something even Tesla has difficulty with.
  • Tesla is a global brand and is very well known but even this company has great difficulty in reaching scale and is still burning billions of dollars in cash.
  • Consequently, Faraday Future should serve as a cautionary tale for electric vehicle start-ups who have got in on the electric and autonomous vehicle hype and now have to live up to reality.
  • I even remain cautious about Tesla because its lack of scale and cash flow could mean that it gets crushed when the big OEMs start producing electric vehicles in big numbers.
  • I remain exceptionally cautious on the Chinese electric vehicle start-ups who have done very little so far and have everything to prove.

Broadcast TV – Sword of Damocles pt. IV.

Reply to this post






Cable TV looks finished while OTA has another chance.

  • The nominations of the Golden Globes have revealed further gains for the OTT players over the traditional content creators also signalling the coming demise of cable TV.
  • The nominations for the 2018 Golden Globe awards have been published where the TV section makes interesting reading.
  • Out of a total 55 nominations, 15 (27%) went to TV shows that were exclusively owned and distributed by streaming services.
  • This is a big increase from last year where 10 (18%) out of 56 nominations went to shows owned by the streaming services.
  • However, 2017 has seen real differences emerge between the streaming players with Netflix pulling away from its competitors Amazon and Hulu.
  • In 2016 Netflix and Amazon were neck and neck on 5 nominations each but this year Netflix has gone up to 9 (behind HBO on 12) while Amazon has fallen to 3.
  • I suspect that this is a reflection of the big increases in content spending that Netflix has made which grew to $6bn in 2017 and $8bn in 2018.
  • However, while Netflix is gaining ground on the traditional content creators, its catalogue is really suffering as other content creators are increasingly pulling their content from its catalogue and going on their own.
  • The latest is Disney which aims to start its own streaming service in 2019.
  • Netflix started life as a distributor of DVDs and was early into streaming, but as it has rightly identified content as the future, it earned the ire of the rest of the entertainment industry.
  • The result could end up being a series of streaming services all with exclusive content from which consumers can pick and choose their subscriptions.
  • To me, this is the death rattle of the cable TV industry.
  • A standard cable TV subscription in 2016 cost on average $103.10 per month (Leichtman research group) for which a large number of channels come as a prepaid package.
  • However, in reality, most users watch only a few of those channels meaning that it if they could subscribe to those channels individually, they would be in a position to save a lot of money.
  • With the content creators all fragmenting into their own streaming services, this is exactly what seems to be happening and I suspect that the amount of money spent on premium TV in USA is going to fall meaningfully.
  • This is likely to be a death sentence for both the cable TV companies and the smaller channels that ride on the coat tails of the big channels.
  • This is because they are receiving income from the all in one subscription payment despite having very few viewers.
  • At the same time there appears to have been a substantial recovery in the number of households making use of OTA (over the air broadcast).
  • According to a Nielsen study commissioned by Ion Media, OTA only households has grown by 41% over the last five years to 15.8m households although this may have slowed significantly since 2015.
  • Furthermore, this is not limited to older generations as the median age of households using OTA and not cable is lower at 34.5 years than the total households using TV at 39.6 years.
  • Although the total number of households switching back to OTA-only may have slowed, there has been real growth in households that also have a fast broadband connection (nScreenMedia).
  • This leads me to believe that users (young and old) are increasingly switching off cable and replacing it with a combination of premium streaming services and OTA TV.
  • This allows the user to have access to a wide range of channels representing almost all the content he was watching on cable at a much lower price.
  • Consequently, while commentators are cautious on the outlook for TV advertising revenues in 2018 and beyond, I think that they could easily witness a recovery having been stalled for some time.
  • I suspect that this is a temporary trend as the spectrum that OTA currently occupies could be re used for much more valuable services as there is no reason why OTA can not be streamed just like everything else.
  • The obvious move is to make the entire selection of channels available on a single, free, ad-supported streaming service.
  • If it is really sharp, OTA will also seek ways to make its offering available in emerging markets which are highly price sensitive and willing to consume advertising in lieu of paying a subscription.
  • Hence, I think that the era of one big subscription is coming to an end and consumers are likely to end up spending less money while still getting exactly the channels that they want without having to indirectly pay for any more.
  • I see cable cutting accelerating significantly in 2018 with real industry change coming in 2020 and beyond.

Apple vs. Spotify – Duck lines

Reply to this post






Two lines of ducks are forming.

  • With a new alliance between Spotify and Tencent and the acquisition of Shazam by Apple, the big players are getting their ducks in a row to dominate the global music industry.

Spotify and Tencent

  • Following on from Tencent’s failure to acquire the company (see here), a compromise has been reached where the two companies will acquire a cross holding in each other and “explore collaboration opportunities”.
  • To me this means two things:
    • First: Tencent will be able to integrate the Spotify service into the ecosystem experience that it assembles to enable it to address developed markets.
    • This could form part of an offering that also has a gaming offering using Supercell and an instant messaging offering using Snapchat.
    • Second: I expect that Tencent will get access to Spotify’s best in class music categorisation and recommendation system for use in making its Chinese service better.
    • For this I expect that Spotify will receive a revenue share from Tencent as I think it very unlikely that Spotify has any real chance of succeeding in China on its own.
    • This is because Tencent’s QQ music is already the leader with 41% share and because like everything else, China’s music market is predominantly about Chinese music for Chinese users.
    • The Chinese music market has been tiny historically, but this is beginning to change as the more affluent end of the market is beginning to pay for streaming services.
    • Consequently, the Chinese market is almost all digital with only a tiny physical presence.
    • It is this change that I think has interested Tencent in Spotify’s technology as technology is what I have long believed underpins Spotify’s superior performance relative to Apple Music enabling it to keep Apple at bay.
  • A close collaboration between Spotify and Tencent could mean a fully global offering with the exception of India which would probably require an acquisition to get a foothold.


  • At the same time, it seems likely that Apple has reached a deal to acquire Shazam, the music recognition company, for $400m.
  • This is well below the $1bn valuation at which Shazam raised money in 2015 but in 2015, there was a much greater supply of belief.
  • The issue that Shazam has had is that it has had great difficulty in making money as 2016 revenues appear to have been around $50m with the company hovering around break-even.
  • I suspect that the company has not grown nearly as quickly as it expected which has meant that profits expected by investors have not materialised making them willing to consider a lower offer.
  • From Apple’s perspective, I do not see this as an acquisition of a service but much more a technology.
  • Shazam has been analysing and recognising music for nearly 20 years and as a result is pretty good at characterising, recognising and understanding music.
  • This is one of the traits that makes Spotify’s service so good as it is able to take that and match it to users’ tastes.
  • Consequently, I see Apple taking Shazam’s technology and incorporating it into Apple Music in a bid to improve its service and compete more aggressively with Spotify.
  • I continue to believe that the best way for Apple to do that would be to introduce a free tier, but that is a whole other discussion (see here).
  • This is great news for SoundHound as the loss of independence of its major rival will make it far more appealing to anyone who competes with Apple at any level.

Take Home Message

  • Streaming has reversed a long decline in music industry revenues and consequently is widely considered likely to become the standard way to distribute music.
  • I still think that the market is big enough for two players to thrive and Apple and Spotify remain at the top and look unlikely to be seriously challenged.
  • Hence, these recent moves look to be aimed at cementing the position of the two leaders ensuring that streaming remains a duopoly outside of China and India.

Qualcomm & Microsoft – Dream with caveats.

Reply to this post






Some space left for this proposition

  • Windows on ARM is back for another crack at the PC market and while I can see a place for it, it is unlikely to cause Intel too much grief.
  • Microsoft and Qualcomm have launched a series of laptops that are running Windows 10 on Qualcomm’s SnapDragon 835 chipset.
  • In addition to offering better battery life, the presence of the chipset enables the always-on functionality and connectivity that users are used to with smartphones and tablets.
  • If that were all that there was to this story, then I would be pretty sure that Qualcomm would quickly take over the PC market but, as always, the devil is in the details.
    • First: Software compatibility: Many of the devices will ship with Windows 10S (for which they are best suited) but will be upgradeable to the full version of Windows 10.
    • Microsoft has compiled Windows 10, Edge and shell to run natively on ARM and had also recompiled a series of DLLs (dynamic link libraries) to ensure that the major desktop applications run properly.
    • For everything else, Microsoft has created an emulator (generally a big drain on performance) that will allow other third-party apps to run with some exceptions.
    • These are: 64bit apps won’t work yet, kernel mode drivers are not supported which means that most antivirus and games that use DRM or anti-cheat software wont work properly.
    • This means that buyers of these devices will not be able to be 100% certain that everything they might want to run will work.
    • I see this as a big sticking point, as failure to perform as expected will infuriate users and create a lot of bad press around these products.
    • This is the same concern that I had around the launch of Windows 10S which I continue to think makes some sense in the classroom but nowhere else (see here).
    • Second: performance. These devices need to perform as well as Intel devices in their pricing tier otherwise buyers are likely to be put off.
    • Given that Intel has much higher gross margins than Qualcomm in silicon, this might be achievable, but it will also depend on the quality of the implementation by the PC makers themselves.
    • Third, market dynamics: The PC market has changed dramatically over the last few years as casual users have deserted the platform.
    • This is because, these users predominantly used a PC for browsing, email and media consumption and smartphones and tablets offer a more convenient and better way to conduct these activities.
    • Consequently, these users have ditched the PCs that they owned and replaced them with smartphones and tablets instead.
    • It is this that I have long believed has been mostly responsible for the softness that has been observed in the PC market over the last 5 years.
    • This trend also means that the users that are left are much more focused on the functions that PCs do really well like content creation and high-end gaming.
    • For these users, performance is critical, and I suspect that Windows 10 on ARM will not be powerful enough for them.
  • This leaves Windows 10 on ARM somewhat in limbo but for students, schools and very price sensitive users, this may represent a good option.
  • Hence, if Intel is going to feel any heat from this, it is going to be at the very low end of the market which is not where it makes most of its money.
  • The amount of traction that these devices get depends mostly on their price and the quality of the implementation by the PC makers but I think that it is pretty clear that the performance driven end of the PC market is almost certain to remain Intel’s

Artificial Intelligence – Zero to hero.

Reply to this post






DeepMind AlphaZero is smarter with 1000x less effort.

  • DeepMind has taken another step forward in the quest for machine intelligence with the demonstration of the rapid training of a single algorithm to play Chess, Go and Shogi.
  • While this is without doubt another step forward, I do not consider that the second major challenge in AI is close to being solved.
  • RFM has identified three main challenges that need to be overcome for AI to really come of age (see here).
  • These problems are:
    • First: the ability to train AIs using much less data than today,
    • Second: the creation of an AI that can take what it has learned from one task and apply it to another and
    • Third: the creation of AI that can build its own models rather than relying on humans to do it.
  • DeepMind’s previous publication took a shot at problem one (see here) and while it represented an advance, I did not consider it to have really solved the problem.
  • Its current publication (see here) takes a shot at problem two, but again has made an advance, but in my opinion, has not really cracked the problem.
  • DeepMind describes a new algorithm called AlphaZero which is a generic version of AlphaGo Zero (Go algorithm (see here)).
  • It uses a deep neural network instead of the specific policy and value neural networks that were designed to play Go in AlphaGo Zero.
  • AlphaZero is then given the rules of Chess, Shogi (Japanese version of Chess) and Go and asked to play itself and to use reinforcement learning to improve.
  • In each case AlphaZero was quickly able to obtain a level of play that allowed it to beat the best algorithm available in each of the three games including the original AlphaGo Zero.
  • It is also highly relevant that AlphaZero did far less “thinking” than its opponents.
  • Each machine was given 1 minute of thinking time and during that time AlphaZero searched 80,000 positions per second for chess and 40,000 per second for Shogi while Stockfish (Chess) searched 70 million per second and Elmo (Shogi) searched 35 million per second.
  • In effect, AlphaZero expended 1000x fewer resources to arrive at a better solution than its opponents due to its use of its deep neural network to tell it where to search.
  • The ramifications for this are substantial as it implies that once trained, algorithms could be easily and efficiently executed on mobile devices where resources remain extremely constrained.
  • However, it is critical to recognise that for each game, DeepMind trained a different instance of AlphaZero.
  • DeepMind started with three instances of AlphaZero which were all identical other than they each had the rules for a different game.
  • However, through playing themselves and reinforcement learning they all diverged from one another as they gained expertise in the specific game they had been asked to play.
  • The end result is that despite a common starting point, the three algorithms become very different by the time that they are capable of playing these games at a very high level.
  • Consequently, to me this does not represent the solution to problem two because one cannot take the Chess version of AlphaZero and have it win at Shogi.
  • However, what it does do is represent a major step forward in the training of algorithms as the AlphaZeros all trained themselves and they all came from a common starting point.
  • This should make training of algorithms in the future easier, quicker and cheaper than they are today which is why this is yet another very significant advance that has been made by DeepMind.
  • Seeing that DeepMind is owned by Google, it is Google Ecosystem devices and services that are likely to benefit from these advances long before anyone does.
  • This will allow Google to differentiate its services more effectively and make them more appealing to users.
  • We gave already seen signs of this where Google is able to do portrait mode with one camera when everyone requires two.
  • This reconfirms my position that it is Google that leads the world in AI developments for digital ecosystems with Baidu and Yandex in 2nd and 3rd
  • Given, Alphabet’s exceptional stock performance this year, Baidu now makes the most interesting and cost-effective entry point for anyone looking to gain exposure to AI.

Google & Amazon – Battle for the smart home pt. V.

Reply to this post






Google has a chance to displace Amazon.

  • I don’t think Google will go out of its way to patch things up with Amazon as having YouTube absent from Amazon devices could disincentivise users from going with Echo products giving Google Home a badly needed boost.
  • Google and Amazon have been sparring for several months but I think that the move by Google to pull YouTube off Amazon ecosystem devices may bring this issue to a head.
  • It is also a demonstration that content is king and as of today, YouTube is amore important platform than Amazon Prime Video.
  • Consequently, I think that Amazon needs YouTube on its devices more than Google does as users will simply go elsewhere to get it.
  • This sparring began three months ago when YouTube pulled its native support from the Amazon Echo Show.
  • This was followed by the removal of Nest products from the Amazon website and the implementation of a clumsy and far from ideal workaround to get YouTube content back on the Echo Show.
  • Google has closed this loophole as of today and will also pull support from Fire TV from Jan 1st
  • This battle between Amazon and Google is peripheral to their core businesses as even in video, they do not really compete directly.
  • YouTube is an encyclopaedia of user generated content while Amazon Prime Video is just like Netflix.
  • Google does have YouTube Red but this is a tiny part of YouTube overall.
  • Consequently, I think this fight is all about the home and here Google is way behind Amazon despite having the better product (see here).
  • This is because Amazon has been much better showing developers love and as a result they have preferred to develop their smart home products for Amazon Alexa.
  • The result has been that almost every device sold will work with Alexa with only a few working with Google.
  • This has changed over the last 6 months but Amazon’s ability to advertise its products on its website as well as giving its cheapest product away for free has allowed it to maintain its lead.
  • With the critical holiday selling season upon us, this is a great time to throw a spanner into Amazon’s works as not working with Google services is going to be a problem for the vast majority of users and may push them to consider Google Home.
  • I still see Google as being on the backfoot when it comes to the smart home but it has closed some of the gap to Amazon in terms of third parties and it remains a superior product.
  • This will be a key battle that is played out in 2018 and the level of support offered by device and service developers when they show their wares at CES in January will be a key indicator.
  • The market remains very lowly penetrated and so there is everything to play for but I still think that in the long-run Google should win as it has the better product.
  • I will revisit this position again once it becomes clear which way smart home developers are inclining for their 2018 product launches.

Xiaomi – Market timing

Reply to this post






Xiaomi considers a big equity event just as growth peaks again.

  • Xiaomi’s genius for timing is confirmed as talk of an IPO is beginning to ramp up just as its recovery growth rates are about to peak.
  • However, this time around, profitability will be open and clear for all to see and it is here where I think the problems will arise.
  • To be fair to Xiaomi, it has executed extremely well and has been rewarded with a period of very rapid growth as its strategy to distribute through methods other than the Internet and to focus on India has paid off in spades.
  • According to Counterpoint, Xiaomi’s performance really turned around in Q2 17A where YoY growth in smartphone shipments went from -8% in Q1 17A to 59% YoY in Q2 17A.
  • This was a result of three big changes implemented by Xiaomi.
    • First: new products. The new flagship Mi 6 launched at the beginning of the quarter was well received and looks to have been the backbone of the recovery.
    • Second: retail channel. I have long been of the opinion that Xiaomi ground to halt because it had fully exhausted the capacity of selling devices over the internet.
    • In order to address a wider slice of the market, Xiaomi has invested heavily in retail with 123 MI stores opened across China and the first results from this push are now being seen in the numbers.
    • Third: India and overseas: Investments in India are beginning to pay off with the Redmi Note 4 becoming the biggest volume smartphone in Q2 17, elevating Xiaomi to No. 2 in India.
    • By far the largest part of Xiaomi’s overseas fan base is to be found in India and this should help the fan base to grow further.
    • However, India can be one of the most fickle markets as it is so price driven and as many Indian brands have found, success can be all too brief.
  • While growth is clearly back at Xiaomi, the comparisons to the torrid time it had in 2016 are really easy as after Q2 2018, growth is likely to slow substantially as the comparisons to the previous year will become much more challenging.
  • Consequently, sometime in H1 2018 is the perfect time to achieve the best possible IPO price as growth will then be at its highest.
  • However, the big question mark for me is profit, as it is through profit alone that an equity based investment can have any value at all.
  • Here, I am still very cautious as Xiaomi’s strategy is based on providing good quality hardware at a great price.
  • This combined with the fact that it does not have Samsung’s scale in handsets means that it is very unlikely to make more than a commodity margin.
  • When I am as kind as I can be to Xiaomi, I can assume that its smartphone ASP is $270 on 118m units shipped in 2018 with $5.4bn in revenues from smart home products.
  • Assuming an EBIT margin of 5%, this gives me 2018 revenues / EBIT of $37.31bn / $1.87bn implying an EV / Sales multiple of 1.3x and EV / EBIT of 26.7x if the company is valued at $50bn.
  • For an EV / Sales valuation, this is not difficult to reach as Apple is trading on 2018 EV / Sales of 2.7x and Xiaomi is growing faster.
  • However, as I have said above, equity valuation is about profit with revenues being used as proxy when there are no profits.
  • Using EBIT, a very different picture emerges as Apple is trading on 8.4x EV / EBIT and at $50bn, Xiaomi would be on 26.7x.
  • Xiaomi is growing faster than Apple but this is unlikely to last very long and its efforts to build a software ecosystem have been crushed by the BATmen at home and are irrelevant overseas.
  • Hence, it has very little with which to differentiate its wares meaning that it has to compete almost entirely on price.
  • Consequently, the most I would be willing to even remotely consider paying for Xiaomi would be double Apple’s EV / EBIT multiple which would give a valuation of $31.3bn, some 37% below the mooted $50bn.
  • To get to $50bn, Xiaomi would need to put up EBIT margins of at least 8.0% which I think is a stretch given the company’s strategy of selling great hardware at good prices.
  • The advantage of an IPO is that these facts will all be laid bare long before investors have to commit to buying the shares.
  • I suspect that its lack of profitability will keep it from going public until it is capable of putting up much bigger profit numbers.



Google – Brain game pt. II.

Reply to this post






Google remains out front in AI but Baidu most interesting. 

  • The first results from Google’s AutoML project are beginning to surface and are implying once again (see here) that machines may end up being better coders than humans.
  • AutoML was announced at Google i/o in May 2017 and failed to attract much attention mainly because I suspect that most commentators did not grasp the significance of the concept.
  • AutoML is neural network that is capable selecting the best from a large group neural networks that are all being trained for a specific task
  • This is potentially a hugely important development as it marks a step forward in the quest to enable the machines to build their own AI models (challenge no. 3 (see below)).
  • Building models today is still a massively time and processor intensive task which is mostly done manually and is very expensive.
  • If machines can build and train their own models, a whole new range of possibilities is opened-up in terms of speed of development as well as the scope tasks that AI can be asked to perform.
  • RFM has highlighted automated model building as one of the major challenges (see here) of AI and if Google is starting to make progress here, it represents a further distancing of Google from its competitors when it comes to AI.
  • In the subsequent months since launch, AutoML has been used to build and manage a computer vision algorithm called NASNet.
  • AutoML has implemented reinforcement learning on NASNet to improve its ability to recognise objects in video streams in real time.
  • When this was tested against industry standards to compare it against other systems, NASNet outperformed every other system available and was marginally better than the best of the rest.
  • I think that this is significant because it is another example of when humans are absent from the training process, the algorithm demonstrates better performance compared to those trained by humans.
  • The previous example is AlphaGo Zero (see here).
  • I see this as a step forward in addressing RFM’s three big challenges of AI (see here) but there remains a very long way to go.
  • These problems are:
    • First: the ability to train AIs using much less data than today,
    • Second: the creation of an AI that can take what it has learned from one task and apply it to another and
    • Third: the creation of AI that can build its own models rather than relying on humans to do it.
  • When I look at the progress that has been made over the last year in AI, I think that Google has continued to distance itself from its competition.
  • Facebook had made some improvements around computer vision, but its overall AI remains so weak that it is being forced to hire 10,000 more humans because its machines are not up to the task (see here).
  • Consequently, I continue to see Google out front followed by Baidu and Yandex with Microsoft, Apple and Amazon making up the middle ground.
  • Facebook remains at the back of the pack and its financial performance next year is going to be hit by its inability to harness machine power.
  • For those looking to invest in AI excellence, Baidu is the place to look as its search business and valuation has been hard hit by Chinese regulation but is now starting to recover.
  • Baidu represents one of the cheapest ways to invest in AI available.