It is not looking good for Huawei, the Chinese electronics powerhouse. Last week, the US government further tightened restrictions on what US companies could ship to Huawei. The US has been steadily throttling Huawei’s ability to purchase parts, and this latest round of restrictions closed a lot of potential loopholes. Put simply, it looks like there is no way for Huawei to acquire any semiconductors or produce its own.
There are really two avenues of the US government’s approach.
First, they have forbidden US companies from exporting chips to Huawei. The latest restrictions add pretty much every Huawei subsidiary and affiliate to the list, so Huawei Europe, for instance cannot buy parts and ship them to Huawei China.
Secondly, the restrictions forbid any company in any country from using any US-created products, including intellectual property (IP), from selling to Huawei. This means that European or Japanese companies are limited in what they can sell. The most prominent, and important, target of this is TSMC, the world’s leading semiconductor manufacturing. The TSMC ban had been in place for a month already, but the latest restrictions make it impossible for some third party, say another Chinese handset maker, from buying parts on Huawei’s behalf.
This cuts Huawei in two ways. First, the straightforward problem that they cannot buy parts for the phones and base stations and routers they build. Second, even if they could cobble together enough non-US sourced parts for their gear, they cannot build their own chips and thus would not be competitive in the market. As we have documented elsewhere, the big electronics companies are increasingly building their own chips for competitive differentiation. Huawei was one of the first to do this in a big way, but now there is no one who can do the physical manufacturing of those chips.
We are repeatedly asked where are the Huawei dependencies. Below is a partial list with an emphasis of parts that have a limited number of suppliers.
Mobile Phones have six key subsystems. Huawei can probably still buy screens, batteries and memory from Asian vendors, who produce their products without US IP. But even here, there may be some obscure tool or IP bloc that will cause the Asian vendors to think twice about supplying Huawei for fear of risking the ire of the US Government, and themselves getting cut off. For modems, applications processors and RF systems, Huawei has almost no options. On the RF front, they can produce some of their own products and buy some of the other parts from Asian producers, but without access to Broadcom’s or Qorvo’s filters, these would be incomplete, rendering the phones non-commercial in most markets. On the baseband/AP side, Huawei essentially has no options – being cut off from Qualcomm, Mediatek and their own products. In theory, they may still be able to acquire modems for lower-priced phones (i.e. 3G/4G) from Chinese vendors, but that is a pretty small market.
Infrastructure: In theory, Huawei may still be able to produce base stations and other networking equipment, but the practicalities are challenging. Their saving grace may rest in the fact that these products do not necessarily need leading edge foundries like TSMC to produce parts. Huawei may be able to make do with 16nm or even 28nm production processes which Chinese foundries can produce. That being said, these products are very complex with thousands of parts. There are probably some distinct pockets within that where Huwaei would need foreign suppliers. For instance, high end network equipment uses specialized TCAM memory, and there is probably no one who would sell these to Huawei.
Intellectual Property: This is a massive stumbling block. On their phones, Huawei has already been cut off from Google’s G-suite of apps, making their phones much less appealing in many markets. More dire is the IP needed to design chips. Huawei cannot access the EDA chip design tools and IP blocs from companies like Cadence and Synopsys, meaning they cannot design their own chips even if they could find someone to build them. This may even include access to Arm’s very essential IP around processors. Without this there is almost nothing Huawei could design. (Admittedly, the drama at Arm China may open a side door to obtaining this IP, but this would be shaky at best.)
So what can Huawei do?
First, they are going to rely on inventory stockpiles that they built up over the past 18 months. However, they have already said they will run out of some of those parts as early as next month. By definition, a limited strategy, especially when it comes to semiconductors which are constantly being obsoleted by Moore’s Law.
Second, they can try to find alternative, especially domestic Chinese, suppliers. The problem is that this is just not feasible. Even though China has 1,300 chip companies, they are by no means comprehensive in their offerings. For example, there are no Chinese RF filter vendors capable of producing all the products Huawei would need. Five years from now, the picture will be very different, but today this is an incomplete strategy.
Third, they can work through third parties. They were clearly working on this path until recently. For example, we believe they were speaking to other Chinese handset vendors about supplying them parts. The latest restrictions make that path impossible. But even before these came into place, we have to question how much any of these companies really want to help their biggest, scariest competitor. Patriotism only goes so far….Given that Huawei is already in trouble in the US for allegedly working through shell companies to sell to Iran, we think there is too much scrutiny on the supply chain for this path to be viable. If some brand new smartphone company suddenly started placing giant orders, the industry would know about it very quickly.
Fourth, a lot depends on how the US government enforces these restrictions. This front is decidedly mixed. When the restrictions were first announced months ago, they quickly began to leak. Every US chip company ran up big legal bills precisely parsing the exact origins of their products, with a surprisingly large number determined to be eligible for export. Other companies were able to continue to supply their China joint ventures (JVs) who then sold to Huawei. Last week’s restrictions seem to cut off that avenue, but there are reports, like this one in the Journal, that claim Qualcomm is trying to get a waiver to continue to sell to Huawei. If Qualcomm is trying, we can assume others are as well. If we had to pick one option, our guess would be this is how Huawei survives, because it comes with a domestic US constituency interested in selling to one of their largest customers.
Fifth, Huawei could split itself up. Spin out the handset and networking businesses into separate entities. This would then shift the onus to the US government to explain how NewCo’s smartphones are a national security threat meriting export bans. And maybe the answer is to split into dozens of entities, further blunting the accusations. Someday we will write a post on the corporate structures Chinese companies employ, but a dozen spin-offs from Huawei is barely the Amateur Tier. The drawback of this is that for it to succeed, each entity would have to declare who their ultimate owners. And that leads back to the heart of the problem. Huawei has drawn scrutiny because the US has claimed that it is owned/controlled by China’s military. Huawei has denied this but has never given a clear picture of who exactly does control the company. The US government would insist on clarity around the ownership of any spin-off, and if Huawei is unwilling to do that for the core company it is unlikely that they would be able/willing to comply for the spin-offs.
None of these are good options for Huawei. Each option is built on a lot of conjecture and big ‘If’s’. We would be very surprised if the company shut down entirely, but the coming months and years are going to be very trying for them.
Pingback: Who benefits from Huawei’s predicament? | DIGITS to DOLLARS·