The U.S.-imposed deadline on Huawei for the more restrictive trade policy took effect on September 15, 2020. By all accounts, Huawei has been expediting parts from all suppliers for some time. Two of the most critical components are the leading-edge semiconductors from TSMC, specifically modems used for smartphone modems and ASICs for radio units (RRUs) mounted on cell towers. Both the smartphone and infrastructure markets require leading edge semiconductors, but smartphone quantities are about 10x that needed for base stations and drive more revenue. However, the choice of what market to support going forward is not so obvious. Given the short time to stockpile, what parts would Huawei prioritize from TSMC?
For 2020, Huawei will ship approximately 180M smartphones supporting a wide range of functionality from low to mid to high tier. Combined, the vast majority will use a modem either designed by HiSilicon (Kirin platform) or by MediaTek (Dimensity or Helio platforms). The demand for 2021 will be relatively flat, so another 180M modems will be needed and all these parts would typically come from TSMC. (Both HiSilicon and Mediatek use the TSMC fab). While not all handsets require the advanced 7nm process, all in the high tier need it in order to achieve the high level computation with long battery life and low heat dissipation. Huawei is the leading handset supplier to the Chinese market at about a 50% share, and it is a 15% player for overall smartphone shipments.
For the infrastructure side, it is a different story. Huawei is a key player, a majority supplier, for 5G base station equipment being rapidly deployed in China. In 2020, Huawei is expected to require about 8.5M processors in base stations increasing to 12M in 2021 (assuming it maintains a similar share). It is straightforward to see that it’s much easier to fully support infrastructure requirements vs. smartphone demand.
Another piece of this puzzle is the supply side. Given that the annual consumption of ASIC-like die (ie. either ASICs for base stations or modems for smartphones) will be roughly 190M each for 202o or 2021, that translates to 15.8M per month, all coming from TSMC. Since Huawei knew a ban was coming, it has been expediting material for the last several months (prior to Sept 15). Huawei already has a large inventory and has stated it can support the forecasted 2020 demand. It is reasonable to assume that the urgent last time buy could have been for any part deemed critical to future operations in the second half of 2020 and through 2021. Working with a 2 month expedite window, there could be roughly 40M parts delivered (15.8M x 2 x 20% upside from expedite efforts).
The final piece is the product impact. On the smartphone side, Huawei has an average sale price of approximately $300 per handset (2019: $67B in handset revenue from 240M handsets sold). As the inventory of 7nm modem die are depleted, this average selling price will be reduced as the production of premium tier handsets will necessarily drop. On the infrastructure side, the numbers are smaller with around $42B derived from ASICs in the Remote Radio Unit (RRU) and Baseband Unit (BBU), due to significant value from software and services to the base station.
Back to what Huawei would order from TSMC. It will consider two key aspects:
- How long until an alternate solution could be available?
- How many complete “kits” can be purchased from suppliers that will be blocked after September 15? (This is important as components that are sourced from non-impacted suppliers don’t need to be stockpiled…saving expense and storage space).
While the handset market clearly is much larger, it makes more sense to stockpile the ASICs used for base stations, and here is why.
- Huawei is not expected to exit the handset business. It will have to step away from the premium segment, which is by all accounts the most profitable, true for any handset maker. Without access to the cutting edge 7nm process, it simply cannot provide competitive performance. This leaves Huawei with a 100M+ handset market in entry and mid-level phones but “only” worth about $25B.
- The network business, in total, was $42B in 2019 with the RRU comprising about $12B of this, and BBUs representing another $10B or so. However, the market growth is 2020 for Huawei is substantial with RRU content double or even triple their previous baseline. There will be a similar lift to other ‘network’ business, so the investment in networks clearly has upside, and loss of the 5G base station endangers the entire $42B portfolio. To be sure, there are a lot of moving parts to account for… Huawei will have to secure everything to make this work.
- The development of a 5G network is a critical mandate in China. Huawei is a key part of this effort and there is no desire on the part of the government to bring in additional vendors to offset a drop-in output from Huawei.
As noted above, Huawei could have as many as 40M parts as an extended horizon stockpile from TSMC. In a best-case scenario for Huawei, this would be about 3 years of material to support base stations. Even this might not be enough to develop a new source for advanced semiconductors but it does allow a substantial amount of the 5G network in China to be deployed, even as Huawei could continue to supply lower tier handsets and continue to develop a Huawei-centric eco-system for handsets, watches etc.
Mobile Experts has been tracking the shipments of many other components used by Huawei in their base station production. They’re still buying radio components in high quantity from Japan, Korea, and Europe. We have seen dramatic slowdowns in shipment by some local Chinese suppliers, so a great deal of uncertainty remains for Huawei’s ongoing production.
Huawei is a critical supplier to the 5G network in China, and while it is also an important maker of premium handsets, there are alternatives. It is unlikely China will want to have more foreign investment into domestic telecom infrastructure and therefore it is most likely that Huawei will support base stations first and handsets second. It is a lose-lose scenario, having to choose one business over the other is difficult and the result in either case is just painful.
Dan McNamara is principal analyst at Mobile Experts, specializing in macro base stations and handsets. His background in RF technology spans more than 29 years in areas of RF design, CATV applications and product marketing for PAs, filters and switches supporting all generations of digital handsets. He is based in Silicon Valley.
"Industry Voices" are opinion columns written by outside contributors—often industry experts or analysts—who are invited to the conversation by FierceWireless staff. They do not represent the opinions of FierceWireless.