5G radio’s introduction in mobile phones definitely adds more complexity to the already complicated RF front-end, and it is likely to exacerbate the phenomenon since price pressure on the RF front-end will be even bigger...
Pushed by 5G megatrend, the RF front-end industry is showing an intense competition. The RF front-end and connectivity markets content increases year over year (YoY) while cost pressure remains important.
Yole Développement announces a 8% CAGR between 2018 and 2025 for the RF front-end market. With US$15 billion in 2018, this industry should reach US$25.8 billion by 2025, confirms the RF electronics team at Yole.
Yole and System Plus Consulting have built an overall and detailed picture of the RF front-end architecture as well as a comprehensive industry outlook. The technical analysis is based on an impressive project including more than 50 phone teardown realized and is now available in a new report, titled RF Front-End Module Comparison 2020 – Volume 1. The report delivers the study of at least 20 FEM and several components found in three smartphones: Apple iPhone 11 Pro, Samsung Galaxy Note10+ and OnePlus 7 Pro 5G and an extensive database of FEM found in 17 smartphones from 2019.
“A deep understanding of the mobile phone’s RF front-end architecture is critical to understand the RF front-end modules market”, asserts Stéphane Elisabeth, Technology & Cost Analyst from System Plus Consulting. From the selected smartphones, System Plus Consulting confirms:
The main phone manufacturers differentiate from each other in the RF field by adopting either an integrated or a discrete approach. In the former segment, market leaders Samsung and Apple, along with smaller OEMs like Sony, LG, Google, and ZTE, are moving towards integration with complex RF modules from Broadcom, Skyworks, Qorvo, Qualcomm, and Murata. “Integrated” players prefer to focus on the user experience with innovative features like “Face ID”, wireless charging, AI camera, gesture recognition, and the human machine interface, thus leaving most of the RF front-end’s complexities to the RF module makers.
Meanwhile, markets challengers like Huawei, Xiaomi, Oppo, and Vivo, which drive as much volume as the market leaders, favor a discrete approach whenever possible. This allows them to keep the RF bill-of-materials (BoM) as low as possible in order to offer aggressive selling prices in this competitive market.
A comparison of teardowns for Huawei’s P20 Pro and Samsung’s S10 illustrates these opposing strategies. The P20 Pro RF board is made of 45 discrete components and four integrated modules (incorporating 25 components), while the Samsung S10 comprises 17 discrete components and eight integrated modules (incorporating 71 components). Consequently, the BoM for the RF front-end in the S10 ends up being double that of the P20 Pro, even though both devices exhibit similar downlink-speed performance. Also, both devices support more than 30 bands and use similar technologies, such as carrier aggregation and 4×4 MIMO.
5G radio’s introduction in mobile phones definitely adds more complexity to the already complicated RF front-end, and it is likely to exacerbate the phenomenon since price pressure on the RF front-end will be even bigger. Indeed, 5G’s penetration rate in smartphones will depend not only on network availability and proposed use-cases, but also on device affordability at the consumer level. Accordingly, we can expect a 30% ratio of discrete components in value to be maintained from 2018-2025.
If LTE and 5G components are cumulated in 5G mobile phones to ensure either the “non-standalone mode’s” operation or for a 5G non-covered area to be connected with LTE, then the question of whether 5G sub-6 GHz and 5G mmWave will both be integrated in the phone is not so straightforward. The cost-over-benefit ratio for the consumer could be too high – plus, a spectrum-holding analysis of the world’s main carriers suggests a divided approach. For example, the first 5G devices were released with different SKUs, depending on whether the carrier integrated a sub-6 GHz radio link or an mmWave radio link. Ultimately, 5G could lead to a regionalization phenomenon instead of being globalized, as was the case earlier for LTE.
Consumer demand for mobile phones weakened in 2018, resulting in a market decline. In this context, the competition has intensified, creating an emulation to push 5G forward. With LTE, RF front-end market growth came from carrier aggregation and MIMO technologies. 5G will augment RF front-end market growth with additional frequency bands, dualconnectivity implementation, and the transition to 4×4 MIMO in the downlink direction, plus a trend towards 2×2 MIMO for the uplink. As a result, the overall RF front-end market (which was $15B in 2018) will grow at an 8% CAGR, reaching $25.8B by 2025. Integrated module growth is forecast at an 8% CAGR from 2018 – 2025, while discrete parts will grow 9% over the same period. Among discrete parts, antenna tuner will grow the most (13% CAGR) because of the combined implementation of higher-frequency bands and 4×4 MIMO, leading to an increasing number of antennae and/or antenna tuners.
The fact that 4×4 MIMO will be mandatory for 5G (at 3.5 GHz and above) will positively impact LNA component growth for either a discrete approach or in-module implementation. And on the mmWave side of 5G, antenna-in-package (AiP) devices have started generating revenue in 2019, with the U.S. being its first target. Yole anticipates this market will reach $1.3B by 2025.
Given the integration of LNA with switches, the industry is moving to 12” RFSOI, thereby limiting growth opportunities for silicon germanium. In the filter domain, the legacy SAW technology will remain stable, while thin-film SAW, BAW, FBAR, IPD, and MLC technologies will see growth opportunities.
In System Plus Consulting’s comparative analysis, Stéphane Elisabeth explains: “From the selected smartphones, most of the components are supplied by Qorvo in number. Moreover, based on the overall devices we identified, the majority is led by the antenna tuner which have 24 % of the share in the function distribution followed by the discrete filter, followed by Murata with 23 %, and Qualcomm with 19 % of the design win mainly due to the discrete filters and the chipset design.”
The last three suppliers, Infineon Technologies, NXP and Wisol, have a very small fraction of the design win share because of their single appliance in the smartphones (Switch/LNA, LNA, Discrete filter).
“The RF front-end leaders still share 81% of the market, with Murata leading ahead of Skyworks and Broadcom”, details Cédric Malaquin, Technology & Market Analyst, RF Devices & Technologies at Yole. And he adds: “Qualcomm, which is already strong in LNA, is catching up along with Qorvo, thanks to the aggregation of TDK Epcos’ filter business. Established companies like Infineon Technologies, Sony, Taiyo Yuden, NXP, and Wisol also possess a market slice.”
These companies generally have manufacturing capabilities for supplying LNA, switches, tuners, and filters, which gives OEMs an alternative to the RF front-end market leaders. Moreover, a variety of fabless companies are emerging, especially in China. Unisoc RDA, Airoha, Richwave, Smarter Micro, Huntersun, and Maxscend are several examples of players scoring more and more design wins amongst the Chinese OEM brands. Obviously, foundries and design houses support this business model for compound semiconductor, silicon, and even acoustic wave filter.