In 2023, it was a year when the domestic RF front-end chip landscape began to take shape. Each segment of the domestic RF front-end industry had its own benchmark enterprises, leading companies, and business camps. Successful domestic RF front-end chip enterprises are all single-point breakthroughs. Finally, five tracks and five benchmarks are formed. It is possible that filters could be a significant market, but discrete filters may face challenges in forming a distinct track or having leading enterprises. The ultimate goal for RF front-end chips lies in module integration. For receiving filters, the trend is towards DiFEM (Digitally Integrated Front-End Module) and LFEM (Low-Frequency Front-End Module), while for transmitting filters, the trend is towards PAMiD (Power Amplifier Module with Integrated Duplexer).
RF front-end chip
Track 1: RF Switch/LNA Benchmarking Enterprise: Zhuo Shengwei
On June 18, 2019, MAXSCEND was listed on the GEM of Shenzhen Stock Exchange. Revenue increased from RMB 44 million in 2014 to RMB 560 million in 2018. In 2018, the revenue of discrete RF switches was 460 million, accounting for 82%; LNA's revenue is 85 million, accounting for 15%. After listing, the revenue in 2019 reached 1.512 billion, an increase of three times. Revenue in 2022 is expected to be 4.167 billion.
After going public, MAXSCEND successfully secured partnerships with various mobile phone brands, positioning itself as the primary supplier and contributing to its rapid growth in performance. At the same time, MAXSCEND moved from discrete switch to DiFEM and LFEM, and its performance grew rapidly again. In 2021, MAXSCEDN also launched PAMiF, officially entering the mobile phone PA, and will inevitably move towards PAMiD.
RF Track 1, led by MAXSCEND and driven by the industry, will change from discrete switch/LNA track to DiFEM and LFEM track, and MAXSCEND is still the benchmark and leader of this track.
Track 2: Phase2 and Phase5NPA Benchmarking Enterprise: VANCHIP
China began to popularize 3G networks on April 1, 2008, and there are three 3G standards: WCDMA, CDMA2000 and TD-SCDMA.
VANCHIP was established in June 2010 and was born for 3GPA. By 2014, VANCHIP decisively gave up 3GPA and made every effort to enter 4GPhase2PA. In 2016, its products began to enter the market, and by 2018, its sales reached 283 million yuan. Since then, VANCHIP revenue has entered the fast lane, with 558 million in 2019, 1.786 billion in 2020 and 3.509 billion in 2021.
In 2022, mobile phone shipments fell sharply. The revenue of VANCHIP in the first three quarters was about 1.777 billion yuan, which is expected to be around 2.5 billion yuan for the whole year, down 30% compared with the revenue in 2021. Nevertheless, Phase 2 and Phase 5NPA products and VANCHIP are still the first suppliers of domestic mobile phone brand customers.
VANCHIP also invested heavily in PAMiD and PAMiF products. In the field of discrete switches, it used to be a secondary focus, but now more manpower and resources are being allocated to prioritize this area. This strategic shift may be seen as an early preparation and layout for VANCHIP to enter the DiFEM and LFEM markets.
Track 3: PAMiD and PAMiF Benchmarking Enterprise: SMARTER MICROELECTRONICS
Founded in November, 2011, SAMRTER MICROELECTRONICS is a company that accumulated earlier in the direction of PA module products, and formed a series of achievements and advantages. The front-end PA/LNA module adopts reconfigurable architecture, with higher integration and fewer wafers, which is helpful for compatibility with larger size filters; It has software tuning characteristics, which is helpful for secondary adaptation after integration; Multi-generation products continue to accumulate, and have the packaging control capability of PAMiD.
SMARTER is also the first mass-produced 5G RF front-end integrated transceiver module L-PAMiF in China. The products cover the new 5GUHB frequency band n77/78/79, and are adopted by OPPO and SAMSUNG mobile phones.
Track 4: WiFiFEM Benchmarking Enterprise: KXCOMTECH
The success of KXCOMTECH and MAXSCEND is an accident for China's RF industry. Indeed, it is both unexpected and reasonable at the same time. Success does not depend on the track, and the market does not depend on the size. Only concentration can succeed. The success of KXCOMTECH and MAXSCEND is reflected in concentration.
Since its establishment in 2014, KXCOMTECH has only done one thing, focusing on WiFiFEM research and development. After starting a business for more than 4 years, the sales volume of RMB 28.57 million was only realized in 2019. 2020 is the first year of WiFi6. KXCOMTECH seized the opportunity of WiFi6FEM and achieved sales of 81.11 million. After proving himself with technology and products, KXCOMTECH achieved 342 million in 2021, which was 321% compared with 2021. Growth; Especially in the context of destocking in the semiconductor industry in 2022, when the performance of most RF front-end chip companies declined, the sales volume reached 203 million in the first half of 2022, which is expected to continue to grow throughout the year. In addition, with the rapid rise of the new technical standard-WiFi7, KXCOMTECH has been recognized by global mainstream manufacturers in WiFi7FEM by virtue of technological breakthrough, and the future can be expected.
Compared with domestic mobile phone PA companies, KXCOMTECH and MAXSCEND have a common bright spot, which is high gross profit. In 2022, gross profit margin of KXCOMTECH is 30%, and the MAXSCEND gross profit margin is 52%.
Track 5: Base Station PA
In the field of base station PA, there are already two listed companies and one pre-IPO company in China. Both listed companies are IDM companies, and the quasi-listed company is fabless Design Company. The market of base station PA fluctuates greatly. In 2021 and 2022, the domestic market size will be about 5 billion yuan, and will drop to 3 billion yuan in 2023. Beside to domestic suppliers, macro base station PA has SUMITOMO Japan and CREE; of the United States; Micro base station PA has SKYWORKS and QORVO.
Technology first, cost is king
Chip Design Company is a technology-based and market-oriented company. Therefore, technology comes first and cost is king. Compared with domestic counterparts, if there is neither technical characteristics nor technical leadership, such a company is a worthless chip design company. The result can only be low-price competition and loss-making sales.
Domestic RF track, market competition has become fierce. While pursuing radio frequency technology, we should also take into account the comparative advantage of cost. Technology is decided by the company's R&D team; The cost is determined by chip designers and suppliers, mainly by suppliers. Different companies, the size difference of chip DIE area is generally no more than 15%, but the price difference given by the same fab will exceed 30%, or even double the difference. Therefore, the chip supply chain determines the fate of chip design companies.
Choice is the key to success and failure. Chip design companies, once they choose the wrong supply chain, will be passive and miserable. Strategic layout of supply chain is the top priority of chip design company. At present, which chip suppliers will decide the fate of domestic RF front-end chips? There are five main types of suppliers:
First. GaAs Wafer Factory (PA)
Gallium arsenide wafer foundries selected for domestic RF front-end chips include WIN, AWSC, SANAN Integrated, LION Electronics, Unicoumpound Semiconduct corporation and CHEMSEMI. Gallium arsenide is the most critical wafer process for RF front-end chips. The performance, stability and price of wafer process are the three most important indicators for RF front-end chip companies, followed by productivity. In the next three years, the production capacity of gallium arsenide wafers will be sufficient, which will lead to more and more obvious lines of chip design companies.
Second, SOI Wafer Factory (Switch&LNA)
SOI wafer has always been a bottleneck for domestic RF companies. SOI wafer suppliers are mainly GF, UMC and TOWERJAZZ, which are big international manufacturers. It is difficult for small start-up RF companies to get good prices and support. If the price is too bad, these RF start-ups can't do it. After several years of development, domestic SOI fabs have made great progress. In another two years, SOI wafers will no longer be a bottleneck. Domestic SOI fabs already have NSEMII, HUAHONGHONGLI and Wuhan XMC.
Third, the substrate factory
The RF substrate is completely localized. The main substrate factories are: ACCESS, SCC, ZDT and FASTPRINT CIRCUIT. Products of ACCESS rank among the top three in the global market share of mobile phone RF chip packaging substrates. ACCESS is the first enterprise in the world to realize mass production of "coreless" packaging substrates by using "copper column layer adding method", and its main products are 3G/4G/5G wireless RF power amplifier front-end modules. By comparison and verification, the performance of Coreless will be better if it is made as PA substrate. SCC now has Shenzhen Factory and Wuxi Factory, and its technological capability is similar to ZDT.
Fourth, the sealing and testing factory
With the increasing integration of RF front-end modules, the cost of packaging and testing is getting higher and higher. The packaging and testing factories of domestic RF front-end chips mainly include CHANGJIANG Electronics Technology, HUATIAN Technology, TONGFU Microelectronics, CARSEM and Ningbo FOREHOPE Electronic. Domestic RF front-end chips have a market scale of 30 billion, of which the scale of packaging and testing is about 6 billion to 8 billion.
V. Filter Factory
In PAMiD products, the cost of filters accounts for about 60% of the total cost. At present, there are mainly three filter suppliers in domestic PAMiD products: MURATA, YUDEN and RF360. Accurately speaking, PAMiD products can't come out without the support of MURATA, YUDEN or RF360 filters.
Each has its own barriers and mutual respect
Although it is also a RF front-end chip, different products have different technologies and different thresholds. Different chips have their own barriers and should respect each other. The following is a brief analysis of their respective barriers:
Mobile phone PA barrier:
In recent years, domestic mobile phone PA technology has advanced by leaps and bounds, and a large number of RF R&D talents have been trained. In addition to PAMiD, other PA products are very mature in technology, and products have changed from technology competition to cost competition. For new startups, the barriers lie in the threshold of big customers and the price of supply chain.
Discrete switch barrier:
The barrier of discrete switches is quantity. Whether a large quantity can be gathered or not, the monthly shipment volume must reach more than 50KKpcs, and then there is a chance to achieve annual sales exceeding RMB 100 million. If you don't get the first supplier of one of the companies of Huami OV brand mobile phones, you won't have the opportunity to achieve hundreds of millions of sales. If you can't achieve hundreds of millions of sales, you can't get a good supplier price.
DiFEM, LFEM Barrier:
The barrier between DiFEM and LFEM lies in SAW filter resources and key customer threshold. Secondly, the cost competition will be extremely fierce. MAXSCEND has the best supply chain cost advantage and the support of major customers, and will undoubtedly occupy the position of the first supplier. There are no less than 20 companies entering this track, so price killing is inevitable. Without a good cost base, products can hardly move in the market.
WiFiFEM Barrier:
Again, mobile phone PA and WiFiPA are different in chip design, and the technical threshold of WiFiPA is very obvious. You can't do WiFiFEM well without time to invest. There will be no overtaking in corners, just by accumulating little by little time. Compared with mobile phone PA, the customer threshold of WiFiFEM is not so obvious, and the cost advantage of supply chain is even less than that of mobile phone PA company, so the barrier of WiFiFEM lies in technology.
Base station PA barrier:
Customers of base station PA are now more likely to choose IDM vendors, a barrier that blocks most fabless chip design companies. Unless there are technological innovations or performance breakthroughs, it is difficult to have another opportunity. The base station PA customer barrier is the highest among all RF front-end chips, just three companies, tow is big company. Other customers can negligible.
Filter Barrier:
The biggest barrier to filters is patents. The filter has a high barrier and can be superior to other RF front-end chips. To do filter, we must do IDM, develop EDA by ourselves, combine design and process closely, and fix packaging technology by ourselves.
Cognition determines thinking, thinking determines behavior, and behavior determines results. While starting a business, we should think about how to improve our cognition of the domestic RF front-end chip track and how to confirm whether our cognition of the domestic RF front-end chip track is correct. We can only compare subjective cognition with objective things and their laws to see if they are consistent. Practice is the only criterion to test whether cognition is correct or not.
Facing the initial pattern, for the founders of RF startups, it is no longer possible to bury their heads in technology. They should look up at the road, improve their cognition of industry pattern, market competition, supply chain, company team, etc., reposition and find a way out, and improve the success rate of entrepreneurship by improving their cognition.
Domestic RF track initially set a barrier to mutual respect for chips-China exportsemi.com
