After meeting with several router clients, I have calmed down. It is not just the Wi-Fi FEM track that is dead, but also the home router market.
Despite the bleak router market, clients are still willing to meet with me for broader discussions. On one hand, many manufacturers that previously promoted Wi-Fi FEM are gradually fading away; on the other hand, only a few companies can keep up with the technology and products. Those who can maintain low prices while keeping gross margins are even fewer. Some are clearing out inventory, while others are transforming. Startups like Sanwu Microelectronics, which are committed to the Wi-Fi FEM space, are basically gone.
The problem is that sticking to it is futile; the linear Wi-Fi FEM track has already been killed. Mobile PA manufacturers are entering the Wi-Fi FEM space at any cost. Even with more than ten Wi-Fi FEM startups exiting, there are still Kangxi and Sanwu Micro, along with 4-5 mobile PA companies involved, and the intensity of competition has not diminished at all. The brutal competition in the market will continue. Faced with a lifeless Wi-Fi FEM track, even if capital has a favorable view and interest in the repetitive chip track, I still maintain my viewpoint: to be value-driven, to act positively and optimistically, while keeping a pessimistic expectation and mindset, and to be responsible to the team and investors.
Many people do not understand why I am so committed to the Wi-Fi FEM track. This may be due to a lack of understanding of hard technology. There are two important points in hard technology entrepreneurship:
1.Choosing the right direction: Some technical directions are pseudo-needs, and the real demand for some technical directions is far less than initially predicted.
2.Long-term technical accumulation: Breakthroughs and improvements in chip technology are actually slow, often requiring 3-5 years. Those companies that seem to be fast either took shortcuts or will lose in the future. The ones that can truly succeed are those who move forward steadily.
The founding team of Sanwu Micro has persisted in the Wi-Fi FEM product for 10 years. If we cannot do this product well, what other products can we do well? Ten years of persistence, 10 years of accumulation; making any chip requires long-term accumulation and a lot of money to navigate many pitfalls. It is unrealistic for us to enter any new chip track now; where will the time and tuition come from?
As long as we are not dead, we will keep pushing forward. In the Wi-Fi FEM track, Sanwu Microelectronics has no way out, moving forward fearlessly. We are increasing investment in technology in the lab and intensifying efforts to lower prices in the market. I can assert again that the market price of Wi-Fi 7 FEM will not exceed 0.2 USD, because there is only one pricing model in the domestic market: competitive pricing. Once there is sufficient competition, there is no bottom line for prices, and the Wi-Fi FEM track also has no price bottom line.
LinearWi-Fi FEM is dead, but non-linear Wi-Fi FEM still has opportunities. Why does non-linear Wi-Fi FEM still have opportunities? Because non-linear Wi-Fi FEM can establish barriers and thresholds; in markets without a moat, there will be bloodshed.
Non-linearWi-Fi FEM (non-linear Wi-Fi front-end module) is a front-end module used for wireless communication (such as Wi-Fi) whose core feature is the inclusion of non-linear power amplifiers (PA) or other non-linear components. The emergence of non-linear Wi-Fi FEM (Front-End Module) is closely related to the development of wireless communication technology and the challenges of hardware design, with the core focus on solving the problem of non-linear distortion under high frequency, high power, and high efficiency demands.
TraditionalPA may enter the non-linear region (such as saturation) at high voltages, leading to signal distortion (such as harmonics and intermodulation interference). However, modern non-linear PA achieves a balance between efficiency and high linearity through design optimizations, such as pre-distortion (DPD) technology. By algorithmically pre-correcting the non-linear distortion of the PA, signal fidelity is improved, which is especially crucial for complex modulations like OFDMA (Wi-Fi 6/7).
As the Wi-Fi standard expands from 2.4GHz (802.11b/g/n) to 5GHz (802.11a/ac) and even to 6GHz (802.11ax/be), the linearity requirements for power amplifiers (PA) and low-noise amplifiers (LNA) become higher. Higher-order modulations (such as 1024-QAM in 802.11ax/be) are extremely sensitive to signal purity, and non-linear distortion can lead to a sharp increase in the bit error rate (BER). The FEM needs to integrate PA, LNA, switches, filters, etc., into a tiny module, making it difficult for traditional linear designs to balance efficiency and performance. When the PA approaches saturation power, gain compression (AM-AM distortion) and phase distortion (AM-PM distortion) occur, leading to harmonic and intermodulation distortion (IMD), for example, the peak-to-average power ratio (PAPR) of OFDM signals forces the PA to operate in the non-linear region. The transient non-linearity introduced by fast switching in TDD systems, along with thermal and memory effects, causes parameter drift due to temperature changes at high power, further exacerbating non-linearity.
The design innovation of non-linear Wi-Fi FEM lies in the digital pre-distortion (DPD) technology, which significantly enhances linear output power (for example, from 20dBm to 28dBm). The emergence of non-linear Wi-Fi FEM is an inevitable result of wireless communication systems approaching Shannon’s limit. Its technical essence is to achieve a balance between efficiency, linearity, and cost by jointly optimizing device physical characteristics (such as materials and processes) and signal processing algorithms (such as DPD, ET). With the evolution of Wi-Fi 7/8, non-linear management will become the core differentiating competitive point in FEM design.
Therefore, without the cooperation and debugging of the Wi-Fi main chip, non-linear Wi-Fi FEM cannot be initiated; without the power curve provided by the Wi-Fi main chip manufacturer, a non-linear Wi-Fi FEM developed in isolation cannot be adapted. More importantly, not all non-linear Wi-Fi FEM can be called non-linear Wi-Fi FEM; as Huawei HiSilicon’s R&D team told me, they hope to reduce the current by 100mA, and reducing it by 80mA is the bottom line.
The path of non-linear Wi-Fi FEM is difficult and has high barriers. It is precisely because this path is difficult yet correct that non-linear Wi-Fi FEM is the future.
END
Finally, I would like to mention that the 21ic Forum (bbs.21ic.com) is recruiting original authors, with a maximum reward of 500 yuan per article. We welcome all netizens to actively submit! 
Click to learn more about the event details.
Previous Highlights:
- Hot search number one! DJI forces employees to leave work at 9 PM
- Breaking! The US announces tariff exemptions for some Chinese products
- Abandoning US citizenship, the chairman of Zhongwei Semiconductor has restored Chinese citizenship!
- Starting work and immediately becoming unemployed! A 14-year-old tech giant in Dongguan announces suspension of operations!
- Star companies explode: power outages, building closures, all employees owe wages, and the boss runs away!
Scan the QR code to follow the video account.
Please click ใโกใ to give the editor a thumbs up.
