The Supply Chain King Laid Low
Since Tim Cook took over as Apple (AAPL) SVP, then COO, then CEO, Apple has mastered the supply chain like no other company. He came in, shut down company-owned factories, and started working with contract manufacturers like Hon Hai. These are amongst the very few partnerships with which Apple is happy.
The result was this astounding chart:
Apple maintains under 10 days of inventory. This is a company selling many millions of devices a year. It is an astounding feat, and as you see, no one comes close to them in this important measure of supply chain efficiency.
But even the king has to go begging sometimes. Apple is still trying to recover from the disruption in their Chinese factories, and their suppliers’ factories back in February. It is an important month every year because it comes on the heels of the lunar new year, so factories are ramping back up after time down. This year, lunar new year shutdown got extended into March. Apple and everyone else have been playing catch up ever since.
The first indication of this was Apple’s Q3, ended in June. Because of work from home, Mac and iPad had unusually good quarters, while the rest flagged a bit. But Mac and iPad would have been even stronger had Apple not had up to 6-week delays on popular models by the end of the quarter.
So that surge got pushed into Q4, just ended and soon to be reported, which will boost the top line. But now is Apple is dealing with these same supply chain issues on their cash cannon, iPhone.
In most years, the iPhone is announced on a Tuesday in the middle of September. Pre-sales begin that Friday, and they usually ship a week after that. So Q4 typically gets a short burst of iPhone sales in that last couple of weeks. This makes Q4 the second biggest quarter in the Apple seasonal calendar in most years. It is also a great bellwether for that year’s iPhone, soon after its release.
But if you’re reading this article, you already know that schedule went out the window this year. The iPhone announcement did not happen until October 13, after the end of Q4. Moreover, the phones are coming out on a staggered schedule:
They are losing over 2 weeks of the quarter for the two more popular models, and 5 weeks on the other two models. That leaves plenty of time for pent up demand and holiday shopping, but it still may mean that iPhone underperforms a bit in Q1 2021, the current quarter.
But let’s begin with the important parts of the iPhone announcement.
This was the big egg everyone was waiting to drop, and there was also the additional announcement from Verizon (VZ) CEO Hans Vestberg that they were turning on their midband 5G network. These midband networks are far more important right now than “true 5G.”
5G is a marketing term, so what it means is in the eye of the marketing executive. The core of it is very high frequency radio signals known as “mmWave” in the industry. But there is a reason all this bandwidth was still available in the 21st Century: these frequencies are very hard to work with.
mmWave only travels short distances, but the even bigger problem is that the signal gets knocked down by just about everything not plastic or glass, even rain or fog. Your hand blocks it, which is very much a problem when the primary use case right now is a handheld device. We’ll talk more about that issue when we discuss Apple’s antenna design, which is one of the key parts to making mmWave actually work.
But all these problems lead to these sorts of maps from Verizon, who is the most aggressive in building out mmWave early in this process:
The dark red is mmWave, what Verizon calls “Ultra Wideband.” Notice that almost 2 years into this buildout, it is still mostly in midtown and lower Manhattan. More importantly, notice how it is only on the streets, not entire blocks except for a few parks. This is because unless you live in a glass house, you will not be getting mmWave indoors. mmWave is hugely complex, is going to take a very long time, and even when it is done, will still not cover everything, even in cities.
Verizon keeps plugging their partnerships with the NFL and NBA to put mmWave in stadiums and arenas. But the details just underline how hard this is. There is no mmWave in the concourses or parking lots because all that concrete knocks it down. But even inside the stadium, the signal is limited to “certain seating areas.” Stadiums are giant spaces that are designed specifically so that every seat has line of sight to numerous places where cells can go. People are mostly stationary, so cell-switching issues don’t come into play. This should be the simplest test case, and why Verizon probably chose it, and they can’t even fill the whole stadium with mmWave.
But then there is the next darkest red in the map, what Verizon calls “5G Nationwide.” This is their midband service, frequencies between LTE, AKA 4G, and mmWave. These are much more manageable than mmWave, and as you see gets far better coverage in the New York map. But even that is very limited right now:
The dark red dots are the cities where they have started rolling out mmWave, but you already saw how limited that was once you zoom in. But also look at how little of the 5G Nationwide they have out there. It is hardly nationwide.
In fact, my neighborhood in LA, just off one of the main commercial streets in the city, does not get Verizon’s midband yet, and that extends about 2 miles down the road.
The basic tradeoff is the higher the frequency, the more data it can carry at much lower latencies. But the higher up you go, the more problems you get with practical coverage issues. So midband is a compromise. It will be a huge upgrade from 4G, but still provide comparable coverage eventually. But it will not be the super fast, low-latency speeds of mmWave.
How Apple Implements 5G
Because of all these issues, making a 5G phone is very hard. The first generation were power hungry heat monsters. When Joanna Stern of the Wall Street Journal tried to test mmWave networks with the Samsung Galaxy 5G in the summer of 2019, she had to carry a cooler with her to keep it from overheating when on mmWave networks, which were even harder to find then.
Everyone’s solution to the power and heat issue is to fall back to 4G whenever possible. But Apple has the advantage here because they control the rest of the stack. They optimized all the software in iOS to run networking more efficiently when on 5G. They also have a new option, “Smart Data Mode,” which will switch back to 4G when 5G speeds are not necessary. For example, you want 5G speeds in Messages when sending or receiving a video, but it is overkill for text. Over a long text conversation, that power savings adds up.
But the bigger issue is antennae design, and this gets into Apple’s troubled relationship with Qualcomm (QCOM). Apple hates being dependent on other companies for key technologies like they are on Qualcomm’s 5G RF chipset, and Qualcomm more than most because of what they consider Qualcomm’s onerous licensing terms. Up until recently, they were suing each other about it. But they are also not that happy with what they are getting. Assuming they are using the X55 chipset debuted earlier this year and not the upcoming X60, this is what Qualcomm was selling in 2020:
Qualcomm has stopped photographing their chips next to pennies which they did for many years, so we can’t tell exactly how big they are, but the reference design requires the X55 chip on the right, plus 3-4 of those antenna modules on the left. They are placed on the edges of the phone, perpendicular to the motherboard. Since your hand blocks the signal, the idea is that no matter how you hold the phone, at least one of the modules will be exposed. Adding it all up, that’s a lot of extra hardware. Here’s what that looked like with last year’s hardware, when Qualcomm was still doing their penny photos:
To-scale image from Ron Amadeo at Ars.
The modem and antennae take up far more room than the SoC. Everything is smaller in 2020, and Qualcomm says that the new QTM525 modules will allow for designs as thin as 8 mm. All the iPhone 12 models are 7.4 mm thick.
The point is that Apple is not looking to put more hardware on their motherboard, they want less. Adding the x50 plus 3-4 antenna modules is a non-starter for Apple.
-Me, in May 2019, explaining why Apple would not have a 5G phone in 2019
So it was obvious even that far back that Apple was going to do their own antenna design, which is a very key component to making a 5G phone usable. They would not even be able to fit Qualcomm’s 2020 modules into the iPhone 12. So this is what they came up with.
Apple gave us only a quick shadowy look at the antenna during the presentation, but that’s it above. There are six segments to the band, probably all with functionality. There are 11 elements all together. I’ve circled in red what I am 99% sure are mmWave antenna modules, and I think the one circled in blue is the third.
The one near the power button uses a little notch on the outside to be able to transmit and receive through the metal band:
It’s unclear what exactly that notch is. Usually, tech reporters would be attending an in-person event, and they would have some brief hands-on time with the phones after the presentation. The possibilities are plastic, glass, or micro laser-cut holes like Apple uses sometimes for speaker grilles.
That module will get covered whenever the phone is held in portrait orientation with the top up, which is the most frequently used orientation.
The other two modules look like they are flush with the glass back, and leverage that to communicate. I’ve circled their rough positions on that last screenshot. So when held in landscape mode, most likely both those get covered up, but the notch gets exposed.
We can be very space efficient, while also supporting the most 5G bands in a single phone.
-Arun Mathias, Apple VP, Wireless Technologies and Ecosystems. iPhone event video.
5G will be supported on over 100 carriers in 30 regions, but that still will leave some disappointed people. All the US phones will support mmWave, but that is not true in every country. Look at the difference between the US and UK iPhone 12:
The US phones have the mmWave notch, but the UK phones do not, because they do not need them. We will know more about all this once they are in the wild and we have some teardowns, which will happen that same day.
One of the interesting things is that this moves Apple farther away from the tapered and rounded edges that Jony Ive loved so much. But these shifts have come from practical considerations. On the iPad Pro and new iPad Air, they are there so the Pencil 2 can magnetically attach and pair. On the new iPhones, they have even more functionality, and it makes 5G possible without a lot of other compromises.
The A14, U1 and LiDAR
Apple spent much more time on 5G than I expected. When they released their first 3G and 4G phones, they talked about it a bit, then moved on to the things they really care about. This is their A14 system on a chip (SoC), U1 positioning and location chip, LiDAR and the camera system. We’ll get to the camera in the next section.
The A14 is the only smartphone SoC using TMSC’s 5nm process right now, and they have scheduled all of TSMC’s 5nm capacity for a while. Scale helps. 5nm means the transistors are 25 silicon atoms wide. This is important because the same sized SoC can support more transistors, almost 40% more than the A13, at higher power efficiency. The CPU design is Apple’s now familiar 4 power efficient cores, and 2 high performance cores, and a 4 core GPU. They claim these are “up to 50%” faster than Qualcomm’s flagship SoC, but provided no comparisons to the A13.
Then there is the Neural Engine, Apple’s machine learning cores, which they have put a huge amount of work into. They doubled the number of cores here to 16, making it “up to 80%” faster. In their own software, Apple is leaning harder and harder into those cores, and they are encouraging developers to also do that for heavy computational processes. It can do 11 trillion operations a second. That sounds like a lot! And they also bumped the speed of the ML accelerators “up to 70%” faster.
A good example of what those cores get used for is the new support for Dolby Vision 4K video. This previously required large cameras and editing suites to produce. Now it can be shot and edited on a device you carry in your pocket. Astounding.
The U1 chip uses a technology called Ultra Wideband, but it is not the same thing as Verizon’s mmWave implementation that goes by the same name. It is a near field location and direction technology. Two devices with a U1 know exactly where the other one is in 3D space. Combine that with all of Apple’s other positioning data, accelerometers, and gyroscopes, and your phone knows exactly where it is in 3D space, and where other U1 devices are as well.
Now add in LiDAR, supported on the most recent iPad Pro, and now the iPhone 12 Pro. LiDAR is a portmanteau of “Light” and “RADAR.” It uses very low power lasers to make a 3D map of the space being pointed at. Put it all together, your phone knows exactly where you are, and what all the objects around you look like in 3D, and also the exact relative positioning and orientation of other U1 devices.
But the very strange thing is that the hardware began getting added in 2019, but to date the applications are still very limited. The U1 makes AirDrop work better in crowded environments, and now enables Handoff to the new U1-equipped HomePod mini. LiDAR is used to make Night Mode photography better, which at least is a major feature.
Apple does not waste space in their hardware, they are the company of 1000 “nos” for every “yes.” That they put these technologies in without a very clear use right now makes me think they have very big plans for this whole positioning/surface-mapping stack I described.
- There are very clear AR implications, and this may all be prepping for the release of the long-rumored Apple AR glasses.
- The also long-rumored Apple location tags.
But my guess is that there is even more here. The U1 is now over a year old in the 2019 iPhones and still does almost nothing. They have big plans for the future in this stack.
Even a cursory look at Apple’s iPhone marketing tells you how important they think the camera is. Along with security and privacy, it has become the major focus of their campaigns. The iPhone 12 brings us 3 new hardware camera systems, but more importantly, more computational power.
iPhone 12 and iPhone 12 mini have the same dual-lens camera system, with “wide” and “ultrawide” lenses. The Pro models have a 3-lens system with the addition of the “telephoto”. The Pro models also have LiDAR, which helps with Night Mode and Portrait Mode. Apple also used the extra space in iPhone 12 Pro Max to give it a bigger sensor, and a longer telephoto lens. That may push some photo enthusiasts to the larger model, even if they prefer a smaller phone.
But the bigger improvements are what’s happening because those camera systems are attached to a very powerful little computer. Apple has spent years focusing on the image signal processor and those Neural Engine cores, and in every iteration of hardware and iOS, they do more heavy lifting. When you snap a photo, your iPhone is taking multiple simultaneous exposures at once, and keeps going if you have Live Photos turned on.
Then the image signal processor and the Neural Engine go to work on it and produce the image you see, all without you ever knowing what just happened. All of that just got better of course, especially on the LiDAR equipped phones. We already discussed one of these under-the-hood improvements, Dolby Vision 4K video. But pros and photo enthusiasts will soon have the control over all these computational adjustments that they have wanted for a long time with the upcoming ProRAW feature. This will give them editing and color grading abilities even beyond traditional digital RAW photography. ProRAW editing is supported in Photos and unspecified upcoming updates to pro photo apps.
Pricing: The Carrier Subsidy Dream Lives
The headline prices in the US are $699 for iPhone 12 mini, $799 for iPhone 12, $999 for iPhone 12 Pro, and $1099 for iPhone 12 Pro Max. But all those prices are wrong, because Verizon and AT&T (T) are hungry to get people on their 5G networks and are willing to pay.
In the first place, the prices of the iPhone 12 and 12 mini are both actually $30 higher. Verizon and AT&T are picking up the $30 for their customers, but T-Mobile (TMUS) customers and those buying SIM-free unlocked phones will be picking that up themselves.
But AT&T and Verizon are going a little farther as well. Both are going to offer fat trade-in value for iPhones, Glalaxys, and Pixels. For example, Verizon is going to give $440 for a 2 year old base model iPhone XS, while Apple only offers $300. AT&T is also offering a 30 month no-interest plan instead of the 24 Apple and Verizon offer, with an $800 trade in for an iPhone 8 or newer in excellent condition. The catch is you are locked in for 30 months and the phone is not unlocked like the other plans.
Confused? Here’s a table with the range of prices you can pay for a base model iPhone 12 if you have a 3-year old iPhone 8 in excellent condition to trade in. These are all interest-free loans.
Free iPhones from AT&T. Come and get it. Apple usually tries to keep SKUs and pricing very simple, but now they have 4 new phones, and 5 different prices on 2 of them, and 4 prices on the others.
Apple has a problem in the smart speaker space, and her name is Siri. Apple does not vacuum up every bit of data like Amazon and Google do. In fact, they make it impossible for themselves to access most of your stuff, to the great chagrin of law enforcement. This is a big focus of their development efforts and marketing, but it is not without a cost, and I’m not just talking about the revenue they could generate from all that user data if they so desired.
The bigger cost is that Siri will always lag Alexa and Google in functionality, because everything she knows about the user that is important is stored on device or encrypted in the cloud. Apple is unable to create the types of algorithms that Amazon and Google do as a result. What Siri is best at, and what Apple emphasizes, is controlling your Apple devices and services. Siri is your servant, but she mostly stays inside the walled garden.
So Apple comes into the smart speaker space with a self-imposed disadvantage: it is not going to be as smart at all the types of things that Alexa and Google are. Apple’s solution to this problem was to focus on the speaker part of the equation, and make a great sounding speaker, but that hasn’t been enough to drive HomePod sales, especially at the price they are charging.
So enter HomePod mini at $99. They miniaturized the speaker part, and it remains to be seen if it matches the impressive sound quality of its full sized sibling, which is also pretty small. I can only tell you that it spurred consumption in this household. We never had an interest in HomePod, but we are planning to get 3 of the minis.
Remember that Wearables, Home, and Accessories is the fastest growing segment at Apple, now eclipsing Services. Since their blockbuster 2015 fiscal year:
They are going to keep going here.
How This Affects Q4 2020 and Q1 2021
Q4 2020 will be reported in a couple of weeks and is in the books. What we know from Apple’s limited guidance:
- iPhone did not get that surge of early orders at the end of the quarter.
- Mac and iPad sales continued their strong work from home charged performance. Q4 is also the back-to-school quarter.
- Services are still mostly App Store and AppleCare. App Store has overperformed in the pandemic, and AppleCare underperformed. AppleCare revenue is partially dependent on in-store service, and many of these were closed. The net was a drag on Services performance in Q3. Apple has also warned that AppleCare has a tough comp with Q4 2019 coming up.
- Wearables, etc. was also mixed in Q3 due to pandemic issues. More people needed wireless headphones for video calls so AirPods did well. However, fashion is a big driver of Watch sales, and without being able to try on in stores, sales there lagged. The introduction of Watch Series 6 may turn that around.
So from that, we get the following assumptions.
- We have no precedent for a no-iPhone September. I took the rate of decline in Q2-Q3 iPhone revenue, 8.8%, and shaved that off of Q3 iPhone revenue.
- Mac and iPad flat with last quarter for work from home, and add a 10% bump for back-to-school for Mac, and 5% for iPad.
- Add 5% QoQ to Wearables, etc. for a couple of weeks of Watch Series 6 sales.
- Services flat with last quarter
Putting that together, revenue would be down 8.3% YoY, $5.3 billion. Q4 is a very tough comp.
This is a drastic reshaping of their income sources, much like Q3. Keep in mind, FY 2019 was hardly a great year for Apple, with iPhone down 14% YoY.
Q4 is harder to nail down because of the iPhone release schedule:
I took an informal poll of smart people I know: will the pent-up iPhone demand be enough to make up for the lost time? The most common response was, “that’s a good question, I don’t know.”
I don’t know either. AT&T and Verizon’s contributions to Apple’s bottom line will certainly help demand. October 23 and November 13 leave plenty of time for holiday sales, but given the still-strained supply chains, the next question is would Apple even be able to meet a flood of demand in such a compressed period? Would their suppliers? They staggered the releases to give themselves a head start on the more popular models, but November 13 to December 25 is a short window.
In either case, I think what this means is that regardless of Q1, Q2 2021 is going to be unusually fat for Apple.
The Short and Medium Term
For years, Apple traded with a P/E in the mid-teens, typical for consumer discretionary. One of my bull arguments from 2010-2019 was that Apple should get something closer to a consumer staple multiple in the low 20s. iPhone is not a staple, but it acts like one, with a much less cyclical revenue pattern than other consumer discretionaries.
Well, I got my wish, and then some.
Apple is trading at a P/E in the mid-30s now, after getting over 40. It is still over double what it was for a decade. If the price were to remain about the same, the P/E will go up after they report Q4, which will likely be worse than 2019. And 2019 was hardly a great year for Apple:
In addition, an elevated stock price means that Apple gets less bang for their buck with buybacks:
The amount of cash the company has been spending has remained pretty stable since FY 2018, but as the share price has soared this year, Apple is getting fewer shares. A third fewer in the TTM.
So, the company will either have to remain on this new slower pace of about 4.6% of shares retired per year, or even slower, or eat into net cash faster. It has chosen a mix. In May, Apple started an accelerated buyback program that got to $6 billion through the end of June, and is slated for $50 billion all together. But again, this won’t be enough to get back to the old pace at the current price.
If your time horizon is short, you should consider taking profits.
The Long Term
That graphic is a header I use in emails and the like. The background is a scan from of the New York Times stock quotes page from September 1982, when my entire meager life savings was 46 shares of Apple stock. I chose that day because Apple was down half a point, and to remind myself daily the Apple also goes down. I sold out when Jobs did, which of course turned out to be a bad idea for the next couple of years.
But I began buying back in 2005, and I have still not sold a single one of those shares. I am a long-term Apple investor is because they are a company built for the long term. Their focus on customer satisfaction is entirely unique in the consumer products business. For years, their focus has been keeping customers happy, efficient operations, and the rest has taken care of itself.
But the bigger reason is their tech stack, now reaching to the sky. We already discussed the location/positioning/surface-mapping slice of this, but it is something that has been brewing since 2007 when Apple popped GPS, gyroscopes and accelerometers inside the first iPhone. Every year they add to the software stack, and some years like the last two, they add hardware as well. In the end, they have a phone that can do what no other phone in the world can do: know exactly where it is in 3D space, and everything that is around it. The possibilities are endless.
Another great example is the A14 SoC. Apple started working on their own SoCs pretty much as soon as the iPhone became a hit. Three years after the first iPhone, they were powering iPhone 4 with their own silicon. At first it was fine that they were as good as Qualcomm’s and Samsung’s, but within a few years, they were clearly much better. Now, it’s just accepted fact that Apple makes the best smartphone SoC in the world, and it only powers iPhone. Now all that work is being leveraged to bring Apple Silicon to Macs.
The point is not just that Apple saves some money, but that they get to make they SoC they want to make, not what Qualcomm or Intel (INTC) wants to make to please a wide range of OEMs. As such, their devices can do things that no other can do, like shoot and edit Dolby Vision 4K on a phone.
Someday, something will replace the touchscreen smartphone as people’s main device. I don’t know when that will be, or what that will be. But I do know this: Apple may not have the first, but they will have the one everyone else winds up copying.
I’ll be back in a couple of weeks after they report.
Disclosure: I am/we are long AAPL. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.