Guys - the teardown of the new Galaxy S7 has been done. For those of us that have debated the stickiness of Skyworks and threats around design in and design out…
Three Skyworks Solutions (SWKS) chips were completely replaced in the GS7.
Looks like Broadcom, Samsung, STMicro, TI and Invensense also lost.
http://www.investors.com/news/technology/samsung-galaxy-s7-q…
That’s got to hurt.
Ant
So they must have known this (obviously) going into last earnings call, yet still gave great guidance/foresight into the back half of 2016. They always beat their conservative guidance. Should we be concerned? I don’t know.
I would think this is already reflected in Skyworks’ sales last quarter. Manufacturing, packaging and shipping of a phone takes time. Certainly they knew it was going to happen a long time ago. Yet still management is optimistic and the company has good earnings. I’m not too worried.
Even if I am correct in this, I would not be surprised if this results in a temporary price drop.
David
I think it’s important not to jump to conclusions. We should verify what the dollar content for SWKS in the Galaxy 7. Just because there are fewer SWKS chips does not mean less SWKS content as some chips are worth several dollars and some are worth pennies. Also, multiple chips can be combined in one chip. So what really happened???
I think there are 2 lines of implications too what has happened…
This isn’t actually the reason why I posted this. It maybe material or it may not and it may have been known or it may not have been known at last guidance stage. In fact they may not even have been allowed to refer to it.
This to me is more concerning. I have always been worried (as I’ve seen it before), how chip sets or component suppliers can be designed in/out abruptly. We have had discussions around how this is possible and how sudden it can take place. This to me is far more worrying.
It seems the competitive strength and moat is not as vast as thought.
If iPhone7 did the same later this year I would assume we would have a real problem on our hands.
Ant
According to chipworks there are no longer any presence (that they could identify) of Skyworks components in the Galaxy S7 Edge:
As would be expected with a worldphone like the S7, there is plenty in its RF section. It is easiest to list them by manufacturer:
Avago AFEM-9040 multiband multimode module
EPCOS D5275 antenna switch module
EPCOS D5287 antenna switch module
Murata FAJ15 front end module
Murata KM5D18098 Wi-Fi module
Qorvo QM78064 high band RF fusion module
Qorvo TQF6260 PA duplexer
Qorvo QM63001A diversity receive module
Qualcomm QFE3100 envelope tracker
Qualcomm QFE2550 digital tuner
Qualcomm WTR4905 transceiver
Qualcomm WTR3925 transceiver
Other parts of note are:
AKM AK09911 compass – we have seen this in over 60 phones!
Bosch BMP280 pressure sensor
DSP D4A1A audio/voice processor
Knowles S1636 microphone
Knowles S1638 microphone
Maxim C551GY5A time-of-flight sensor
Maxim MAX77854 PMIC
Maxim MAX77838 PMIC
Maxim MAX98506BEWV audio amplifier
Qualcomm WCD9335 audio codec
Qualcomm PM8996 PMIC
Qualcomm PM8004 PMIC
Samsung C3 image processor
Samsung S2MPB02 PMIC – seen in multiple Samsung phones
Samsung S6SA552X touch controller
STMicroelectronics L2G2IS gyroscope
http://www.chipworks.com/about-chipworks/overview/blog/samsu…
In the Galaxy S6 there were two identified modules:
Design Wins
We’ve been able to identify multiple chips inside the Galaxy S6:
Samsung Exynos 7420 SoC
Samsung K3RG3G30MM-DGCH 3Gb LPDDR4 SDRAM and Samsung KLUBG4G1BD 32GB NAND Flash
Samsung Shannon 333 Modem, Shannon 533 PMIC, Samsung S2MPS15 PMIC, Samsung Shannon 928 RF Transceiver and Samsung Shannon 710 Envelope Tracking IC
Broadcom BCM4773 GNSS Location Hub
InvenSense MPU-6500 Gyro + Accelerometer
Skyworks SKY78042 Multimode Multiband (MMMB) Front-End Module (FEM)
Avago AFEM-9020 PAM and Avago ACPM-7007 PAM
Samsung C2N8B6 Image Processor
Maxim MAX98505 Class DG Audio Amplifier and Maxim MAX77843 Companion PMIC
Samsung Electro-Mechanics 3853B5 Wi-Fi Module
N5DDPS2 (Likely Samsung NFC Controller (P/N needs to be confirmed)
Wolfson WM1840 Audio CODEC
Texas Instruments BQ51221 Single Chip Wireless Power Receiver
Skyworks SKY13415 Antenna Switch
STMicro FT6BH Touch Screen Controller
https://www.chipworks.com/about-chipworks/overview/blog/insi…
Sincerely,
Charlie
It seems the competitive strength and moat is not as vast as thought.
Given the purported history of SWKS being involved in designs a couple of years in advance, this teardown can hardly be news to Skyworks. “Oh, surprise, this phone we just released has none of your chips in it” is hardly going to happen. Aside from a current teardown being ancient history in terms of design wins or losses, the question is, what actually happened. Was a simpler SWKS set of chips replaced by a more complex one, potentially replacing chips from multiple vendors? Did someone else’s chips replace the SWKS chips and why? Was the goal going for cheaper cost even at the expense of features and, if so, does this relate to a trend?
I haven’t seen anything yet that provides this context.
So, three SWKS chips were REPLACED by products from some other vendors. And no other chips (individual ones or integrated chip set) from SWKS could be identified.
I thought SWKS worked with major OEMs to incorporate the its design in long before the product release. Guess Samsung is not one of them, at least for Galaxy S7.
Fortunately, Galaxy S7 had lukewarm review so far.
http://wccftech.com/galaxy-s7-manhattan-scores/
Zangwei
Any chance the tear down is not accurate?
From SWKS last CC:
Vivek Arya
Dave, beyond the sort of well-advertised slowdown or inventory adjustment at your largest customer, what has been the trends at your Korean and your Chinese customer? I think they’re all trying to assess whether this was a one quarter problem for your industry or whether there are any other longer-term impacts. If you could give us some insights into that, I think will be very useful for investors.
David Aldrich
I think, as Liam commented a moment ago, Vivek, the China TAM story and the upgrade to 4G has been very positive for the company. In fact, I think we were something like 20% year-over-year growth in our China revenue in the December quarter. And that’s content, that share that’s being lined out with the right set of baseband and SoC partners.
**And if you look at our customer, Samsung, on our second largest customer, we’ve seen very high attach rate now with our more complex solutions in the Galaxy 7 and then in their emerging 16 platforms.**And the fact, Vivek, is our China customers and Samsung they are all adopting, virtually across the board, a more complex system, RF system that’s highly integrated for size, for current consumption, and just to merely handle the complexity and get these designs, these SKUs, out into the market on time. And that’s been very, very good for us; very bad for the discrete companies; very, very good for us.
Seems hard to believe that SWKS wouldn’t know 6 weeks ago that they were left out of G7…
Brian
Wow! So do we beleive Aldrich or whoever did the tear-down? Looks like they can’t both be telling the truth. Aldrich has a pretty solid track record, I don’t know who did the tear down or what motive they might have for false reporting. Seems impossible that they could have just overlooked SWKS in the new Galaxy.
I don’t know who did the tear down or what motive they might have for false reporting.
I don’t think there is any motivation in false reporting, it just seems a matter of a fact type report. In is not like they single out Skyworks at all.
I would believe they are just trying to be accurate though it is possible they made a mistake.
Also possible that they have duel supplier for the same phone and they just happened to tear down a phone without Skyworks? Not sure if they would do that often but didn’t Apple so it it recently and there was a big reaction that the battery would drain faster if you had the phone with a Samsung chip vs who ever the other supplier was?
I briefly worked for a company that supplied some parts that went into Nokia phones back when they were still popular. The company didn’t get designed into every phone, they always kept 2 or 3 suppliers that could do what they needed and used them on different models. That way they would have more leverage when introducing a new model. If you had delivery issues, quality issues and quoting to high, they could use one of the other suppliers for the next model.
I briefly worked for a company that supplied some parts that went into Nokia phones back when they were still popular. The company didn’t get designed into every phone, they always kept 2 or 3 suppliers that could do what they needed and used them on different models. That way they would have more leverage when introducing a new model. If you had delivery issues, quality issues and quoting to high, they could use one of the other suppliers for the next model.
JDC, This could very well be the case. I was thinking the same thing. But wouldn’t that still be a blow to our thesis that other chip suppliers couldn’t make SOCs with the same expertise and complexity that SWKS could, right?
Matt
Long SWKS
MasterCard (MA) Ticker Guide
But wouldn’t that still be a blow to our thesis that other chip suppliers couldn’t make SOCs with the same expertise and complexity that SWKS could, right?
Yeh I think that is the point Matt. I would forget about rear looking points about who knew what etc and think about the implications.
However neither the sudden design in/design out nor possible regional variety of like for like content doesn’t say much about the Skyworks’ complexity advantage or moat.
Ant
wouldn’t that still be a blow to our thesis that other chip suppliers couldn’t make SOCs with the same expertise and complexity that SWKS could, right?
Seems to me all of you worry too much. This is one data point from a customer which is not critical to Skyworks’ business. Samsung could have chosen Qualcomm/TDK products for any number of reasons that have nothing to do with the validity of your investment thesis.
Customers come and go in any business. Skyworks is no different no matter how good their product. The important fact is the overall trend, not the individual data points. If my thesis is wrong that Skyworks has a superior product a pattern will form of customers moving to other products and SWKS earnings/revenue decreasing. I don’t see that happening. SWKS management is also very optimistic and open about their business. That doesn’t sound like a company worried about loosing customers.
Maybe I’m wrong but for now this seems just an interesting fact to keep in mind for the future.
I half way agree, but it concerns me that Aldrich responded in the way he did to a direct question about SWKS business outlook with their “Korean customers”. That is a very thinly veiled question about Samsung.
I’ve always felt that Aldrich was a straight shooter. He did not appear to play games, so this casts some doubt on his trustworthiness.
A basic investment thesis of mine (I think shared by Saul) is that if you can’t trust management find a different investment opportunity. So, I’m a little nervous right now. I’m very long SWKS. Market gyrations have not bothered me (well, OK, not bothered very much). This gives me pause . . .
It seems the competitive strength and moat is not as vast as thought.
It always raises my hackles when Aldrich says that SWKS is the best at what it does with very little competition. He is very confident in that, which is a large part of the investment thesis. At the same time he is shouting “hey, we have big margins and few players over here” which seemed likely to, and apparently has, attracted competition.
I’m inclined to think they will outgrow this until it shows up in earnings, but it certainly bears watching.
4wheel.
Samsung could have chosen Qualcomm/TDK products for any number of reasons that have nothing to do with the validity of your investment thesis.
With all due respect this is exactly the point of the thesis. With companies in possession of a strong moat the point is that customers cannot just chop and change suppliers they are tied to a provider.
The point isn’t how Skyworks is doing with sheer weight of customers it is about nor whether there is choice or not it is about how much stickiness customers have with Skyworks. Clearly not as much as we had hoped for given the ability to cut out Skyworks in one version release during the Smart phone LTE wave that Skyworks is supposed to be dominant in.
Ant
With all due respect this is exactly the point of the thesis. With companies in possession of a strong moat the point is that customers cannot just chop and change suppliers they are tied to a provider.
The point isn’t how Skyworks is doing with sheer weight of customers it is about nor whether there is choice or not it is about how much stickiness customers have with Skyworks. Clearly not as much as we had hoped for given the ability to cut out Skyworks in one version release during the Smart phone LTE wave that Skyworks is supposed to be dominant in.
Ant
I stand corrected. My investment thesis is apparently quite different from yours. My apologies.
We might do and that’s totally ok Othalan. It depends how much a moat is part of one’s investment thesis.
A lot of what Saul and others have discussed and debated about has been around the moat thesis. There hasn’t been any real debate around Skyworks commercial results and performance.
I guess the reason this is so important is that many of us have seen so often semiconductor companies burn bright and then collapse due to substitutability and lack of a differentiated moat.
At the same time Skyworks has claimed a competitive and differentiated advantage in its complexity.
Ant
SWKS management is also very optimistic and open about their business. That doesn’t sound like a company worried about loosing customers.
I wouldn’t expect management to be anything different until they have to. I haven’t followed this company closely but I do recall a lot of the discussion is that they have a large moat as they are doing things in the analog domain where everybody else is digital and not so easy to be replaced.
They maybe great at what they do doesn’t mean they can’t be easily replaced by somebody like Qualcomm. I own AMBA but I have similar worry that they can easily get replaced even if they have better products no matter how optimistic their management is