Huawei and the Global Supply Chain
Huawei and every other hardware company in the world exist inside a global supply chain that has been maturing for a couple of decades now. Like everyone else, they source parts and IP from the US, China, Korea, and Europe. Any break in this supply chain can cause problems ranging from the annoying to the existential.
Huawei has been blacklisted by the US Commerce Department, and US firms can no longer legally do business with them. This has also extended to several non-US companies, most importantly ARM, who uses enough US IP that they decided to play it safe and suspend contacts with Huawei, a long-time customer.
There is an existential threat not just to Huawei, but also to their US suppliers, especially the smaller ones who rely on Huawei for a large portion of their revenue. There is a huge opportunity for Korean companies with a lot of IP and products like Samsung (OTC:SSNLF). If they are able to remain a supplier despite extreme US pressure, they will be able to pick up some of the business that Micron (MU), Qualcomm (QCOM), and Broadcom (AVGO) are losing.
This is a fluid situation. While I believe that this was a key inflection point in the trade war with China, it may turn out to be only a minor skirmish, like the ZTE ban. But until it is resolved, I would avoid the entire US semiconductor sector (XSD).
Huawei Was Ready/Not Ready for This
Huawei has engaged in a lot of hand waving in the past couple of weeks regarding their Commerce Department blacklisting. They claim that while it will slow them down, it will not stop them. But the headlines are coming quickly, and today they didn't even seem sure when their Android fork would be available in China. They are feverishly stroking non-US suppliers in an attempt to get them to ignore all this, but the global nature of the supply chain will remain a hinderance.
They are obviously shoveling as fast it gets thrown at them. But beginning with the phone, I wanted to see what kind of prospects Huawei has if the ban continues past the 90-day waiver period.
The short answer is that Huawei seems to have been preparing for some for some sort of autarkic existence for some time. Either that or they are total control freaks like Apple (AAPL). If you look at progressive year-after-year teardowns of their P-series flagship, fewer and fewer parts are sourced elsewhere.
Many of the parts they used to get from Qualcomm and others are now sourced internally through their HiSilicon division. This same division is going to have to work overtime in the coming months and years.
The first roadblock for which there is no easy solution is Google cutting them off. They still have access to the open source parts of Android, and claim they have been working on an Android fork for some time now. There was a bit of confusion here, but they say the new operating system, likely called Ark OS, will be available in China in 2019 and elsewhere in 2020. The other issue is apps, and we'll get into that as well.
The second major roadblock is being cut off by ARM, which licenses the IP for the cores in Huawei's CPU/GPU system on a chip (SoC). ARM is a British company that is now owned by Japanese SoftBank, but their lawyers have concluded that they use US IP that is covered by the ban. They have not been more forthcoming than this. Their relationship with Huawei goes back a long way, and they are trying to walk a fine line here.
Without these cores, Huawei cannot build SoCs for their phones, but the problem is much deeper. Much of their entire product line uses ARM cores. If this holds up long-term, it is even a bigger blow than Google.
Google and Android
When a phone manufacturer licenses Android from Google (GOOG) (NASDAQ:GOOGL), they actually get three things, and only the first one is free and open source. Ron Amadeo at Ars helpfully compiled the 3-layer stack into a graphic:
Huawei will still be able to access all the basic core libraries in the first column (though not the Android logo), but the second two columns will be off-limits. All those APIs in the second column are super-important, because Android apps use those libraries for basic functions, like maps and location.
As you can see, the third column has much that is easily replaceable, like calculators, contacts and calendars, but there are also irreplaceable ones like YouTube, Gmail, Drive/Docs/Sheets/Slides, Maps, Google Now, and of course the Play Store.
Huawei claims they have been working on their own OS, likely an Android fork, for some time now. If true, I have to believe they thought there was a non-zero chance that Google would cut them off one day, and this day has come. But even if they had already begun preparing, the task ahead is huge.
Alaa Elshimy, Managing Director and Vice President of Huawei Enterprise Business Group Middle East, told TechRadar:
Huawei knew this was coming and was preparing. The OS was ready in January 2018 and this was our 'Plan B'. We did not want to bring the OS to the market as we had a strong relationship with Google and others and did not want to ruin the relationship. Now, we are rolling it out next month.
Elshimy turned out to be wrong: Huawei still has no release date for the OS beyond 2019 in China, 2020 elsewhere. So there is some confusion here. Elshimy also claimed to TechRadar that Android apps would work out of the box with the new operating system, likely named "Ark OS" or similar.
Like Elshimy's first claim of a June release, this seems highly unlikely to me. Just the maps and location issues alone will require many apps to recode to whatever Huawei provides as a replacement to Google's services. Here's some other things Huawei will have to code, buy or license to produce a modern smartphone.
- Ad network
- Auto integration
- Contactless pay
- Wearables/Fitness APIs
- Photo syncing service and app
- Cloud storage and backup to it
- Games APIs
- DRM APIs
- Calculator app
- Calendar app
- Camera app
- Contacts app
- Email Client
- Messaging client
- Printing APIs
- Office apps
- Phone app
- Books/Movies/Music/TV/Magazines stores
- App store
- Voice assistant
Some of these, like the app store, Huawei already has, but the rest they will need to scramble for, and that is not a recipe for success, especially when everyone knows they are a motivated buyer.
Moreover, apps from US developers will not be showing up on the Huawei AppGallery store. According to the most recent tally from App Annie, 7 of the top 20 apps in the Chinese Google Play Store are from Google. 3 are from Facebook (FB) or Twitter (TWTR). 2 are from European developers, and the remaining 8 are all VPNs. Half the top 20 will not be available, including 6 of the top 8.
Other apps in the top 50 that will not be available:
- Steam and Steam Chat
They have a tough slog ahead of them here, but they have proven over the years to be patient, and that they keep plugging away.
The place where Huawei was really prepared, but also most screwed, is on the hardware side. Based on SystemPlus Consulting's teardown of the Huawei P30 Pro, I came up with this list of suppliers
Starting with the items in green, these are mostly items that Huawei's HiSilicon division designs and fabricates through TSMC (TSM), who will remain a vendor. As you can see, they've been busy at work. If I made this chart for the P10 Pro, from just 2 years ago, there would be a lot more Qualcomm and others in the mix. Huawei still uses Qualcomm Snapdragon SoCs for their lower end phones, but they will have to use their own going forward.
A key thing to note is that for most of those HiSilicon chips, I couldn't determine what, if any, third-party chipsets they were licensing. The Hi6422 has a Broadcom chipset, but this is likely replaceable in an SMPS. But if there are other chipsets that are not, some of those green items go to red.
The biggest issue is with the Korean firms which are under tremendous pressure both from the US and China. Which way they decide to go will be key here. They are not just important current suppliers, but moreover will be the sources for some of the replacement hardware, like storage and DRAM. If the Korean companies decide to cut off Huawei, it will make all this much more difficult, and again turn green to red.
In the blue items, these are mostly issues dealing with the international standards organizations. They have already been cut off by the WiFi, Bluetooth and SD associations, which means they get no support, logos, and can't use those terms in marketing. They look to be working on proprietary solutions, but that is hugely suboptimal. So they will forge ahead with their own standards, but not having "Bluetooth" or "WiFi" may be a deal-killer for many buyers.
The final blue item is NXP (NXPI) which makes the best-in-class chip for NFC contactless payments. They are Dutch, but have stopped shipments to Huawei from their 3 US factories. They have not revealed what chips they are, but likely something can be worked out here, since they have facilities elsewhere.
The less important of the two red items is Gorilla Glass from Corning (OC), which clads just about every mid to top end phone around. Moreover, wireless charging requires glass on the back, and 5G radio chips will require even more glass on the edges, a point of weakness. They will have to choose a cheaper and less durable option like Dragontail from Asahi (OTCPK:ASGLF). Asahi hasn't published drop-testing in years, which tells you what you need to know there.
But the most important thing in all this, even more so than Android, is ARM cutting off Huawei. ARM Cortex CPU cores and Mali GPU cores are the basis of not just Huawei's phones, but pretty much their entire product lineup that isn't using Intel or AMD CPUs, which they are also cut off from. This includes much of their network, server and 5G hardware which they are looking to for future growth.
This is an existential crisis if not resolved.
We only have BBC's reporting and a brief, very uninformative statement from ARM to go on. But it seems as if Huawei will be allowed to use whatever it was using up until May 22 when ARM suspended all contacts. This means that their current chip lineup is safe, and likely the next generation, but after that, they are on their own.
There are open-source projects available as a replacement for ARM cores:
Amber is an ARM-compatible core design. But it is only 32-bit and uses an instruction set from the 1990s.
MIPS is a non-compatible design that is in the process of being open-sourced. So far, only the low power 32-bit cores have been open-sourced, and the timetable for the rest is unclear. However, Google did port Android to MIPS through version 4, so it is not undoable. Joel Hruska at ExtremeTech summed up the massive scale of the task:
Huawei would likely find itself in the unenviable position of bringing up its own Android fork on the ARM-based SoCs it can still sell under license while simultaneously porting its OS to MIPS, fleshing out the capabilities of its own Android fork to compete with Google (something only Amazon has really tried to date), and building a brand-new SoC on a new instruction set with an all-new toolchain. Then it needs to optimize said OS to run on MIPS, to ensure it can offer equal or superior performance to its older ARM hardware. In our theoretical example, this would be made easier by the fact that Huawei is stuck on whatever ARM IP it licensed in 2019, while its MIPS processor would reflect the performance characteristics of 2023-2025 silicon.
It would take years to pull all the pieces of this puzzle together. It reliably takes 4-5 years to design a new CPU architecture, and while mobile designs can be simpler than an AMD or Intel chip, Huawei would be truly starting from scratch.
Finally, RISC-V is a widely used architecture for small, low power embedded chips. It is also not up to speed of current ARM designs, and opens up the same issues as MIPS.
This is a monumental boulder they have to push uphill.
The Flip Side
Eek! I used to be a Lumentum (LITE) bull, but no more. NeoPhotonics (NPTN) is pretty much done if it holds. This is going to leave a large hole in many US income statements, including some of the most important US chip suppliers. Of the bigger names, Qualcomm and Broadcom take the biggest hits.
Conclusions: An Existential Moment
This all may be cleared up before the 90-day waiver ends, and become quickly forgotten like the ZTE ban. But this has a different flavor to it. The Chinese have shifted into 5000-Year Mode, and I don't think it is just rhetoric.
I wrote this at the end of last year, but it still holds in my mind:
I continue to agree with Jack Ma that the trade war will be measured in years, not months.
On the US side, Trump has absorbed the teachings of outcast-amongst-economists, Peter Navarro, and they are determined to bring China to its knees. I am not sure what will steer them from this course. Trump's biggest problem here is that he has never read Sun Tzu:
If you know yourself but not the enemy, for every victory gained you will also suffer a defeat.
Trump engages in mirroring, which is that he assumes everyone else is motivated by the same things that motivate him. This served him fine when dealing with NY real estate types and post-Soviet investors, but it serves him very poorly with the Chinese.
On the Chinese side, I'm going to guess that by contrast, Xi has read Sun Tzu and spends a lot of time trying to understand US motivations. Xi does not have midterms or a re-election to worry about, and the Party has already shown in Tiananmen that when push literally comes to shove, the People's Army has little regard for the people. The pressure on him from below is minimal and not an existential threat to the Party, which is what they care the most about.
The Chinese view themselves as the center of the world, the Middle Kingdom, do not respond well to this sort of bullying, and are too big to bring to their knees. Moreover, this is a country that, just within living memory, has survived:
- The Rape of Nanking and other atrocities of Japanese occupation
- Civil War and the Communist Revolution
- The Great Leap Forward
- The Cultural Revolution
- The Death of Mao and the succession chaos that followed
- The ascension of Deng, the move towards a quasi-capitalist economy, and all the social dislocation that accompanied that.
- Tiananmen Square
I don't think they are scared of tariffs.
The threat to Huawei and their US suppliers is real. To Huawei, this is an existential moment, and if they cannot find a way to regain access to Google's services stack and ARM's cores, the road ahead is uphill and rocky. There is no guarantee of success, and many ways for the company to sink.
For their US suppliers, if this doesn't turn around soon, it has already blown a hole in their Q2 guidance, and their Q3 projections are going to disappoint. I would avoid the entire semiconductor sector until this sorts out. Poor NeoPhotonics.
But the broadest effect of this all will be to bifurcate the global supply chain in technology, which has been growing more interconnected until now. The Korean companies' dilemma highlights this. They just want to sell hardware and IP to whomever has the money to buy them. They are being forced to choose now, and both choices are terrible from their perspective.
This will either dissipate into a minor skirmish in the war like the ZTE incident, or become a turning point that has massive consequences for the future of the global economy. Unfortunately, I'm betting on the latter.
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.
Editor's Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.