Huawei – Nowhere to run pt. XXIV.

Re-inventing the silicon wheel makes no sense.

• Huawei appears to be planning on building its own semiconductor fab that is independent of US technology in a move that looks certain to fail.
• I think it would be better off using SMIC which despite its problems, is better at making semiconductors than Huawei is.
• Huawei seems to be intending to build a fab in Shanghai that uses no US technology.
• It intends to source equipment from Chinese companies like Naura and AMEC both of whom sell etching and vapour deposition equipment as well as second-hand equipment that is already in the hands of non-US companies.
• The plan is to start on the 15-year-old technology at 45nm, make its way to 28nm by the end of 2021 and reach 20nm by the end of 2022.
• To me, this strategy makes no sense at all.
  • First, smartphones: Even if Huawei were to succeed in this ambitious timetable, it would not help the smartphone business at all.
  • This is because silicon for competitive smartphones is currently being made on 7nm or 10nm with Apple being the first to go to 5nm with the A14 Bionic launched last month.
  • In order to manufacture at these geometries, one needs new equipment some of which is only available from US companies.
  • Hence, while Huawei remains on the entity list, there is almost no chance that it will be able to get any of this equipment.
  • Second, basestations: Being able to manufacture at 20nm would allow Huawei to make chips for its basestations, but given that most if competitors currently use 14nm, its equipment would be disadvantaged compared to the equipment of Ericsson, Samsung or Nokia.
  • Huawei’s 5G solution is widely recognised to be the best at the moment but Ericsson is rapidly closing the gap and Ericsson also has re-committed Nokia hot on its heels.
  • This superiority may make up for some of the silicon-derived shortcomings, but this will only help in the home market as international operators are increasingly spurning Huawei.
  • Third, SMIC: which is also been made subject to export restrictions like Huawei.
  • I can’t see any reason why Huawei would start from scratch in semiconductor manufacturing when it would be able to use SMIC.
  • SMIC has quality, yield and sourcing issues for sure and its ability to go beyond 14nm is now highly questionable but it has been around for 20 years and certainly knows more about making silicon chips than Huawei does.
  • Huawei is good at designing chips to be made by others (TSMC), but this has not been possible since September 15^th when TSMC halted all shipments to Huawei.
  • Consequently, I think Huawei would get up and running much more quickly if it went into a partnership with SMIC rather than starting from scratch itself.
  • This is unlikely to help much with its basestation or smartphones business, but I suspect it would be much better than starting from scratch on its own.
• The fact that Huawei is even considering this move is an indication of the seriousness of the situation that it is now faced with.
• Its executive team has done a superb job of mitigating the impact of US sanctions for over a year but now the cupboard is bare (see here).
• This means that when its current stockpile of silicon chips runs out, it will be unable to make a range of its products meaning that its market share in those products both at home and overseas may plunge to 0%.
• This will have most of its impact in 2021, creating further opportunities for its rivals (Ericsson, Xiaomi, Samsung and Nokia).
• Of these, the one with the most upside is definitely Nokia.
• Nokia’s shares fell nearly 20% after the new CEO kitchen-sinked expectations and was very straight with the market about the difficulties that the company now faces and what has to be done.
• I think this no-nonsense approach is exactly what Nokia needs to have a good chance of a turnaround, and I am far happier with the new CEO than the market appears to be.
• Hence, I think the opportunity in Nokia has just increased in size but investors will need to be patient as this turnaround was always going to take some time.
• I have a position in Nokia with a 2-3 year time horizon and I am inclined to add some more at these levels.