Year after year, the elimination competition in China's chip design industry continues, and the landscape remains unsettled.
Last week, I visited a networking client who mentioned that three years ago, at least a dozen companies manufacturing WiFi FEMs (Front-End Modules) approached him. This year, only six or seven remain. My assessment is that in the future, there will be no more than four domestic WiFi FEM manufacturers in the networking router market.
Similarly, the head of RF R&D at another leading domestic networking enterprise expressed the pain of selecting WiFi FEM suppliers, unsure of which companies will ultimately survive. Listed companies have become a crucial reference for their selection.
Only by continuously interpreting the market and clients can a company know where it should head. In the router WiFi FEM product segment, only a few listed companies and Sanwu Micro remain. It remains to be seen whether Sanwu Micro can create another miracle.
Therefore, chip companies all want to go public, as an IPO can solve two problems: funding and branding. Funding helps companies survive and increase R&D investment; branding helps companies penetrate previously inaccessible clients and markets, while also enhancing product premium.
Without the SSE STAR Market, it would be difficult for chip companies to go public. When every chip track is crowded with listed companies, listing on the SSE STAR Market becomes equally challenging.
In the future, domestic chip companies will be divided into only two categories: listed and unlisted. Unlisted companies are not necessarily inferior to listed ones; they may even operate better and eventually reach a point where going public or staying private is entirely their own choice. Chip companies that are unprofitable, rely solely on primary market financing, and cannot go public are outside the scope of our discussion.
However, unlisted chip companies must also solve the same two problems: funding and branding. Unable to raise money from the secondary market through an IPO, they must learn to make money from the market. At the same time, they need strong brand awareness, treating brand building as equally important as making money.
What is a brand? A brand is trust. Trust has become the highest cost in the market. Some chip startups do not focus on making good products; they only aim to quickly launch products to generate sales revenue and then tell stories to secure financing. This leads to mediocre performance, poor consistency, and unguaranteed quality. After causing numerous issues for clients, customers are increasingly afraid to adopt products from chip startups.
In the early days, working at overseas chip companies, there was a natural sense of trust when promoting products. Clients dared to try any new product and even competed to be the first or among the first batch to adopt them. It is not that overseas companies' products never had issues, but problems were extremely rare, and when they occurred, both parties would actively resolve them.
When it comes to domestic chips, clients first doubt your performance, consistency, and quality. You need to first list which clients you have served, how much you have shipped, and provide proof. Therefore, the first step for domestic chips is extremely difficult.
Without establishing brand and market trust, unlisted chip companies will be in a very passive position and dare not innovate. It may take half a year or even a year to make progress in pushing innovative products to the market, whereas a listed company can achieve it in six months to a year, seizing your market opportunities at a faster pace.
How can unlisted chip companies build brand and market trust? First, look at what clients care about most. Here, clients are divided into three tiers:
1. Major brand clients
2. Small and medium-sized brand clients
3. White-label clients
Let's start with white-label clients. They care most about price, followed by performance, and lastly quality. Why do they not care about the delivery capability that brand clients value? Because they never discuss delivery, only spot goods, and have no demand planning. Playing in the white-label market is a different chip game; low prices and spot availability are always the prerequisites, followed by performance and quality.
Many domestic chip startups grew from the white-label market, where the primitive accumulation of chip trial-and-error and shipping scale is completed. Many shipments here are either at zero gross margin or negative gross margin, subsidized by capital.
Those who have been in the white-label market for a long time operate with a different chip design philosophy and supply chain system: mediocre performance, mediocre quality, and always spot goods.
What I want to do is serve major brand clients and small-to-medium brand clients, selecting the first-tier supply chain system in the industry, with performance and quality placed first. As for delivery, these brand clients will provide demand plans, and orders are relatively stable with rarely drastic fluctuations, making the delivery challenge much smaller.
The first step in building brand trust and influence is to focus on a niche market. Achieving the number one position in a specific niche market is the beginning of brand building.
Being number one in a niche market does not mean a brand image and reputation have been established. It may simply mean seizing the first-mover advantage, where performance and price have a first-mover edge, which cannot sustain long-term market position.
A true brand is trust. Clients believe that choosing this chip brand is the best decision, making it the optimal choice comprehensively in terms of performance, price, quality (consistency and reliability), delivery, and supply chain resources.
Quality is the most difficult to evaluate in advance in the chip industry. Clients believe that only after cooperating for a year or even longer can they truly assess quality. Companies like Huawei worry the least about chip quality issues because they have a quality evaluation system; products must pass their tests to be adopted. This is how Huawei has become the best endorsement in the chip industry.
Regarding chip quality, grasping the following key points will prevent delivery quality issues:
1. Performance must have margins to leave room for consistency. Here, consistency refers to both chip consistency and the consistency of the client's PCB (Printed Circuit Board) assemblies.
2. Focus on PPM (Parts Per Million). Consumer chips generally range from 200 to 500 PPM, with 200 PPM being relatively good and 500 PPM being the minimum requirement. I require Sanwu Micro's chips to be at 200 PPM. Based on my experience, the primary focus is on having high yield requirements; I set the mass production test target at over 99%. Secondly, test controls must be strict, and controls cannot be relaxed for the sake of yield.
3. CPK (Process Capability Index) metrics. In chip mass production, a CPK ≥ 1.33 is the universally recognized passing line, with critical processes usually requiring ≥ 1.67. CPK measures the production line's ability to stably produce qualified products. It can be simply understood as a comprehensive score of 'accuracy' (how close the average is to the specification center) and 'stability' (how small the fluctuation is). It also reflects the convergence of test data metrics, which is related to both chip design and process fluctuations.
Starting this year, I will focus on highlighting to clients that in the seven years since Sanwu Micro was founded, there has not been a single quality issue or delivery problem. Industry information circulates; if I tell a lie, my reputation will be ruined, and clients will no longer trust the Sanwu Micro brand. A brand is trust, and reputation is the brand.
Just last month, I made a decision to abandon the mobile phone WiFi FEM product line and the drone PA (Power Amplifier) product line, going all out to focus on the IoT (Internet of Things) market. We will focus on IoT FEMs, RF switches, and LNAs (Low Noise Amplifiers), aiming to have the most comprehensive product line, the best performance, the best quality, and the most timely delivery in the IoT RF front-end market.
However, Sanwu Micro has not given up on router WiFi FEMs. R&D for router WiFi FEMs has never stopped, and we are going all out to develop WiFi 8 FEMs.
Compared to pursuing an IPO, I prefer to pursue building a branded chip company. When Sanwu Micro is mentioned, clients feel a sense of trust—trusting Sanwu Micro's performance and quality, as well as its delivery capability and long-term operational capacity.
To achieve this goal, I have proposed two '1+1' strategies: the IoT market plus the router market, and the domestic market plus the overseas market. Sanwu Micro has completed the layout for these two '1+1' strategies and is continuously advancing them.
I want to build a profitable chip company with brand value. Perhaps in a few years, chip technology will no longer be the greatest value of a chip company.