We usually do an annual review of the market for RF semis around Mobile World Congress. That is not happening now, and probably not this year at all. Since we touched on the subject earlier in the week with a look at Qualcomm, and we have had a few questions about the overall RF landscape, we want to expand a little on what is in store for the other RF competitors.
The first question we get asked is – what about Broadcom? Broadcom makes a lot of money from selling high grade filters to the leading phone makers. We tweeted yesterday that Broadcom’s (then known as Avago) purchase of Infineon’s filter business in 2008 for €21 million was probably the best technology M&A deal in history, at least in terms of cash-on-cash returns. Their position is unlikely to change.
Phones, especially high-end phones, still need to a lot of filters. Qualcomm does have some filter capacity, but that seems to still be lagging the turnaround of the rest of their RF business. We think that may change, but it is too soon to tell. Further out, it is possible that if Qualcomm really delivers on the vision of a unified RF front-end, tightly coupled to the phone’s compute capabilities, they could design phones with lower filter requirements. This is what we called the Holy Grail. It is absolutely possible, but it is not easy, not going to happen soon, and will not entirely eliminate the need for filters. It may happen, just not today. Broadcom is still going to make a lot of money selling filters for the foreseeable future.
The story for Murata, another leading filter vendor, is largely similar. They are probably a bit more exposed to Qualcomm’s filter capabilities, but again nothing seems likely to change on this front soon.
So the next question, what will happen to the leading incumbent RF active component vendors – Skyworks and Qorvo. These companies are starting to look very exposed, even if this fact is not reflected in their share prices yet. These companies both now find themselves with a new competitor in their core market, and it is a competitor with a history of steadily rolling over every other player in mobile silicon. On a standalone basis, both companies have better products and operations than Qualcomm, but factor in Qualcomm’s sheer scale and ability to sell a complete solution (please do not call it a bundle) and the outlook looks considerably bleaker. There are already rumors that Qualcomm has taken away RF sockets from Skyworks for this year’s iPhone, a major win if true.
Skyworks best protection at this point is the fact they tend to run a very tight operation. Qorvo is less solid operationally, but does have a better filter business. They are the junior partner with Broadcom in an effective duopoly for high-grade filters. That being said, neither Skyworks nor Qorvo have much left in the way of competitive barriers. Their best hope for years has been poor execution at Qualcomm and customers’ reluctance to further reliance on Qualcomm. The former has changed dramatically and the latter is not much upon which to rest a serious defense. We think it is unlikely that both companies make it through the decade as independent companies.
Further over the horizon, there are two other factors to consider. The first is new technologies from publicly-traded start-ups Resonant and Akoustis. [Note: we own a position in Resonant.] Both companies have promising technology which could alter some of the economics of the business, but neither is large enough to deliver this on their own.
The second factor – and a perennial in all matters of semis – is China. There are a number of emergent Chinese RF companies. Most of them operate at the lower rungs of the market for active components (i.e. not filters). They do not alter the overall landscape much today, but that could change. Also unclear is the fate of Huawei’s RF business, which we believe was starting to becoming quite capable. That is likely to get dragged down by the company’s broader problems, but might make the difference under other ownership. Finally, we continue to search for Chinese filter vendors. We believe they are out there, but they remain very low profile. For now.