Since 2022, newly listed chip stocks on the Sci-Tech Innovation Board have frequently faced share price declines. Among the 17 listed chip stocks, 11 have experienced a drop, accounting for 60%. Netizens summarized three reasons: first, these companies are facing losses and poor profitability; second, there is a significant gap in technology compared to similar foreign companies; third, the market size of the sectors these companies are in is small, limiting future growth potential. The rational pricing of chip companies in the secondary market has also led to cautious investment in the primary market.
Recently, I attended a meeting for investors in the semiconductor industry where we discussed current issues in semiconductor investment. Some viewpoints suggest that the frequent price declines of semiconductor companies indicate that the current financing bubble in semiconductors has peaked and will undergo a process of deflation. However, this does not mean a decline for the entire industry; rather, it suggests that valuable companies will stand out and gain more opportunities, which is beneficial for industry development.
What constitutes a valuable chip company? It is, of course, a company that can make money for its shareholders in the future. However, in reality, many chip companies regard getting their chips to light up as the goal of starting their business, similar to the previous trend of launching apps during the app startup boom. A few years ago, some college students developed an app on Android, released it on the app store, and then inflated the user numbers with their peers to deceive investors into providing hundreds of thousands in angel investments. Now, with the mobile internet boom over, most of those apps have disappeared in the waves of time.
Making chips is significantly more challenging than developing apps. Let’s discuss how difficult it is to run a chip company.
First, there is the challenge of team building.
An app can often be developed by a single person, but a chip cannot be completed by one person alone. For a small digital chip, for example, you need 3 designers, 2 verifiers, and 1 backend engineer, which means at least 6 people. This is merely the minimum staffing requirement to ensure the normal operation of chip development; however, many companies struggle to meet even this basic configuration, let alone a large-scale chip development team.
Once the team is assembled, does that solve all the problems? Not at all. The stability and management of the team is a significant issue. Regarding talent stability, most startups experience high turnover rates. In the startups I know, there is usually a turnover of personnel every 2 to 3 years. Additionally, management issues arise because people from different backgrounds have different working styles, and it takes a long time to refine the team’s working model into something stable.
Second, there is the challenge of technology.
Currently, many chip startups exist, but few truly develop from 0 to 1. Many companies rely on previous experiences to purchase a bunch of IP to piece together a chip. When compared to similar competing products, they gain a slight advantage by stacking SRAM, adjusting bus widths, tuning clocks, and tweaking timing. Is this the core technology of the team? These companies merely verify their IP products by repeatedly completing tape-outs under the guidance of IP companies. Therefore, it’s hard to see where their technological moat lies. Some companies aspire to create high-standard products, such as high-speed IPs, but are hindered by a lack of domestic talent in this area. Of course, some domestic chip companies perform exceptionally well, such as those achieving high precision in ADC/DAC and low static power consumption. These companies have accumulated significant technical expertise in specific fields over a long time, not merely by purchasing a few IPs and hiring a few experienced engineers.
Third, there is the challenge of tape-out.
Does modifying purchased IP guarantee a successful tape-out? Just because the first tape-out is successful, does that mean the second will be? Not necessarily. We conducted a survey and found that 30% of engineers have experienced tape-out failures. The reasons for tape-out failures are diverse, with one of the funniest being that a certain chip manufacturer dropped their wafer on the floor when retrieving it from the factory, damaging it. Thus, while a significant portion of tape-out probability is controllable, some aspects are beyond control.
If the process is well executed and risk management is in place, achieving a successful tape-out is not particularly difficult. A few years ago, a successful tape-out by a single team was a rare event, but today, such occurrences are commonplace, and tape-out failures have become industry news.
Fourth, there is the challenge of mass production.
Successfully tape-out chips and ensuring they work are two different matters. Real products have high demands for chip power consumption and reliability. Some electronic products are sold globally, with significant variations in humidity and temperature; a product that works well in coastal areas may fail in high-altitude regions, likely due to issues with chip reliability. Some electronic products have poor battery life, leading users to mock them as hand warmers, which indicates poor power management. Some electronic products frequently crash, and the causes could be varied: low chip yield or bugs within the chip. Additionally, some chips require extensive certification tests before being integrated into products, such as automotive chips needing AEC-Q100 certification.
Even if mass production of chips is not an issue and reliability tests have been passed, does that guarantee success? Many chips are close to mass production in large companies but are canceled due to market reasons. Therefore, the fifth challenge in chip entrepreneurship is market challenges.
Fifth, there are market challenges for chips.
I have seen some chip companies that have been established for many years and have developed several chips, but no one uses them, a problem that is even more pronounced in early chip entrepreneurship.
There are two types of situations here: one is companies that have no customers and are attempting to drive new products to market with their chips. After developing certain chips, they find no market demand and must search for it themselves. The other type has clear competing products and aims to replace them in the market, but the market does not respond. Taking digital chips as an example, the sales issues encountered include that the performance of their chips is not significantly better than competitors, and switching chips poses risks for some companies; the development environment of their chips is not compatible with that of foreign competitors, leading to high learning costs and long development cycles; the documentation for chip development is insufficient, leaving them unsure how to resolve issues; and foreign similar chips may be cheaper. For example, with certain EDA tools, sales issues arise when engineers are already accustomed to using foreign counterparts, and the domestic products come with learning costs; EDA tools are not used frequently, and when the need arises, they may still choose foreign products; and the learning resources for domestic EDA tools are inadequate.
Therefore, the market is a significant issue; no matter how excellent the chips are or how sophisticated the technology is, if no one uses them, the result is still a financial loss.
Once the chips open up the market, does that mean the company is saved? Such thinking is naive. There are further challenges ahead.
Sixth, there is the challenge of chip production capacity.
Currently, both mature production lines and advanced processes are facing capacity constraints, leading many to joke that whoever controls capacity controls the printing press. Many companies have been unable to deliver chips on time due to production capacity issues, resulting in lost orders. At this point, the resources of chip companies are put to the test, so is it feasible to start a chip company without some resources? Absolutely not.
Seventh, there is the challenge of chip costs.
Does selling chips guarantee profit? Not necessarily. Many chip companies invest heavily in R&D without considering costs, yet the selling prices of chips do not meet expectations; even with high sales volume, they still cannot cover R&D and production costs. If a price war breaks out over chips, the result is that everyone ends up losing money. Many companies, including large ones, are developing high-end chips due to their strong bargaining power. But how many startups can create high-end chips? Is it difficult for chip companies to achieve profitability? Extremely difficult. After overcoming the seven challenges mentioned, a company may only then have a chance at profitability. Just when a company believes it can sustain profitability, it may be mistaken.
Eighth, there is the challenge from competitors.
If you are lucky enough to develop a chip that becomes a bestseller in the market and earn a lot of money, can others just sit idly by? Chinese companies are not good at original creation but excel at imitation. For instance, this article has been published before, and within no time, several media outlets plagiarized it without my permission. It’s frustrating, of course, but I don’t rely on writing for a living. However, many chip companies are severely impacted by imitation. Previously, a certain company developed a chip, and when others saw its market success, they rushed to develop similar products, resulting in that company’s subsequent products failing to sell well. Foreign countries protect intellectual property rights very well; that’s why the functional modules of chips are called IP, which stands for intellectual property core, highlighting the importance of IP protection abroad. Even if the code is public, most chip companies dare not use it.
Ninth, there is the challenge of cyclical changes in the chip market.
The last challenge is the cyclical changes in the chip market. Analyzing the revenues of major chip companies reveals that the chip market experiences cyclical changes. Each cycle represents a market reshuffling. The reasons for these cyclical changes are tied to shifts in consumer products; for instance, fingerprint recognition was very popular a few years ago, leading to a surge in fingerprint chip development. Nowadays, Bluetooth wireless headphones are trending, resulting in a spike in headphone chip sales. The cyclical changes in the market bring new opportunities while squeezing out old markets. Chip companies that fail to adapt during these transitions face the risk of elimination.
So, does successfully tape-out a chip mean the company has succeeded? Sorry, lighting up the chip is merely the beginning of chip entrepreneurship; there is still a long way to go to profitability.
E课网(www.eecourse.com) is a professional integrated circuit education platform under Moore Elite, dedicated to cultivating high-quality professional talent in the semiconductor industry. The platform is guided by the job requirements of integrated circuit companies, providing a practical training platform that aligns with corporate environments. Through both online and offline training methods, it rapidly develops students to meet corporate needs.
E课网 boasts a mature training platform, a comprehensive course system, and a strong faculty. It has planned 168 premium semiconductor courses covering the entire integrated circuit industry chain and has established 4 offline training bases. To date, it has deeply trained a total of 15,367 individuals, directly supplying 4,476 professional talents to the industry. It has established deep cooperative relationships with 143 universities and has organized 240 corporate-specific IC training sessions.