The Radio Frequency Front-End (RFFE) semiconductor market has enjoyed steady growth. Increased number of shipments was an early driver back in LTE days which was followed by an explosion of global frequency bands and corresponding increase in RF content to support it all.
Over time, five RFFE suppliers have managed to dominate the semiconductor supply, accounting for 85% of all content. Four of the these are U.S.-based: Broadcom, Qorvo, Qualcomm and Skyworks with Murata being the fifth. Each of these have a roughly equal share with the annual numbers bouncing around a couple of percentage points depending on the design wins in major flagship phones for a given year. All the details of the RFFE market share are available in the latest RFFE report from Mobile Experts, found here.
On the smartphone side of the equation, the top six OEMs are Samsung, Huawei, Apple, Vivo, Oppo and Xiaomi, representing 75% of all smartphone sales, with China brands accounting for 40% of the smartphone total. Without U.S.-based RF suppliers, China smartphone OEMs have a gap for RFFE components.
The challenge is that the smartphones designed today require a high level of integration that has thus far only been demonstrated by the top 5 suppliers, making them key to all smartphone OEM planning. The constant pace of handset releases does not allow for wholesale replacement of the Complete Front End, CFE, (also known as LlPAMiD) modules without some compromises. After the initial trade ban imposed by the U.S. government on Huawei in mid-2019, there was a significant drop in U.S. content in subsequent handsets. Most notable changes were the many Huawei-branded RF components along with an increase in content from Murata.
Huawei has a huge R&D budget, $15 billion in 2019, but that must cover a lot of activities other than smartphones…like telecom infrastructure. For comparison, the total R&D budgets of Qorvo, Skyworks, Broadcom and Qualcomm for 2019 were on the order of $11 billion with approximately $2 billion to $3 billion going towards components for smartphone development. The rub is that this annual R&D expense benefits greatly from the cumulative learnings over many, many years. In addition, all the RFFE suppliers have in-house filter fabrication technology, a key enabler and differentiator in complex smartphone designs.
Huawei would have to invest heavily in both design and fabrication infrastructure to close this gap. It is a lot of money but also time, measured in years…building factories and getting a stable process can be expedited but there is a limit. Power Amplifier technology has some flexibility as foundries such as WIN Semiconductor are already significant suppliers to RFFE modules. But, as with TSMC, there could be a lot of political pressure brought to bear (from both sides of this equation) that would make this more than a little uncertain as a secure option for the China market.
The beneficiaries, at least in the near-term, could be Japanese and European semiconductor suppliers. Murata has all the components to create all the various types of RF modules needed in 5G smartphones and can support all the players in the China smartphone market. Murata is a leader in diversity modules that integrate filters, switches and LNAs, often in very small form-factors. Stepping up from diversity modules to CFE modules is achievable and something that Murata has been working on for several years. Companies such as Infineon and NXP could increase small signal RF component sales.
Splitting the RFFE supply chain would have a negative impact on all sides. Four of the top five suppliers would lose access to a market that drives more than 350 million smartphones per year. This number is just the China market. Of course, the China OEMs would aggressively promote handsets in other markets such as greater Asia, Africa and South America that could account for another 150 to 250 million smartphones that would not have U.S. RF content. R&D spending from current U.S. suppliers would likely stay the same to maintain the constant improvement demanded by the market but the return on those investments would be decreased, creating a more challenging business environment. One rough estimate is that the loss of 350 to 500 million phones from the “available market” would reduce gross margins by about five percentage points for the American suppliers.
For Huawei, the initial push to replace the lost supply chain would be billions of dollars per year, for several years, just to fill the gaps. In addition, the RF supply eco-system in China would need some sort of investment to get pushed up the learning curve. We estimate that Huawei would have to increase RF-only development spend at least $10 billion over the next two years, to create competent GaAs fabs on the mainland for GaAs as well as BAW or thin-film SAW technologies and integration design capability. Assuming China continues to supply 350 to 500 million phones per year over the next 10 years, this surge of R&D spending would add a staggering $10 or more in cost to every phone.
Note the above costs do not include the problem of CMOS ICs manufactured by TSMC. That’s another crash R&D project to ‘catch up’ on CMOS fabrication below 10 nm….another huge investment of capital that would add $10 or more of R&D cost to every phone.
Of course, there is no guarantee that a surge of R&D spending in China would achieve the same level of performance as the American companies. Today, Huawei is dealing with the loss of highly complex filters (quadplexers and hexaplexers, etc.) by decreasing functionality in the handset, and dropping some Carrier Aggregation combinations from their phone line-up.
It is yet to be seen what solutions Huawei will bring to the market as the more restrictive trade rules come into play later this summer. It is likely we will see more Huawei branded RF as well as more Murata modules. The longer this requirement for separate supply chains continues, the more Huawei and China Inc will be forced to invest to close the supply gap. It would be a net loss for the entire smartphone industry.
Dan McNamara is principal analyst at Mobile Experts. His opinions are his own.