What's still holding back China's huge fabless potential
08/01/2004
The industry's current "up" cycle has been notable because of its strength despite the lack of a new killer application; the cycle has been buoyed instead by many minor new applications and in large part by the existence of a "killer geography" — China.
If all things were equal, Chinese fabless companies would have great advantages over those in the US, Europe, and even Taiwan in terms of proximity to end markets and manufacturing prowess.
But all is not equal. China's advantages, while real, currently are not being realized. For example, even the tax policies that many say give domestic companies an unfair advantage are not important to the majority of Chinese chip suppliers that are not yet profitable.
The great gasp
The market expectation for the emergence of a strong semiconductor industry in China can be measured by the global gasp heard when China announced its intent to "stabilize" its economic growth. For the past 25 years, China has achieved an economic miracle with average GDP growth above 8% annually. However, as GDP has become the main standard to evaluate government's performance, local officials have turned a blind eye to development in other fields, including medical care, education, culture, and environmental protection. Ma Kai, minister of the State Development and Reform Commission, said in late May that the government was considering slowing the country's GDP growth rate to 7% this year to cultivate a "scientific approach" to improve social development.
In effect, China's central government doesn't want the mainland's booming economy to overheat. However, it's hard to imagine government correctly navigating a complex economy where even the slightest miscalculation affects the global economy.
Semiconductor executives are understandably anxious about the impact of this policy on both near- and long-term plans. China is a very important market for fabless companies since many of them are already highly dependent upon the strength and growth of the Asian region. China presents huge end-market opportunities and will increasingly become a significant manufacturing engine for companies. The Chinese market also offers a large reservoir of potential design talent, and it is both a potential competitor and an ally for suppliers worldwide.
As the semiconductor manufacturing supply chain matures in China and more successful companies emerge, such as Semiconductor Manufacturing International Corp. (SMIC) with its recent initial public offering, China is establishing a well-functioning ecosystem. We have seen other geographies like Taiwan use this cluster effect to catapult themselves and become driving forces in the industry. System houses in China also play a particularly important role with their own brand names and sales channels. They will have strong power in defining product specifications and standards, particularly for consumer and communications products such as digital TV, 3G phones, and broadband services. Thus, companies that are close to the system houses, and can understand and meet their requirements, will succeed.
How fabless companies can succeed in China
 Percentage of fabless companies by geography. (Source: Fabless Semiconductor Association) |
Many credible organizations estimate there are 400 fabless companies in China. The Fabless Semiconductor Association (FSA), however, believes this number is exaggerated by a minimum of 50%. Major foundries indicate there are closer to 70 legitimate fabless companies that are potential customers in China. FSA believes that as many as 50 fabless companies will require 0.13µm and below technologies in the next 12 months, indicating that if the number is actually closer to 100, China will be a major force in the market over time. China currently represents about 7% of fabless companies, about half of the number based in Taiwan (see figure).
Once some of these challenges are addressed, Chinese fabless semiconductor companies will make an impact on the global market. Challenges include experienced staff shortages, technology deficiencies, and insufficient financial support:
Lack of design talent maturity. Although China boasts that 7000 chip designers were employed in 2003, most were inexperienced and unable to make a significant impact on a design team, according to iSuppli Corp. Design engineers need at least five years under the right tutelage to be fully functioning design team members. In the meantime, most companies are focused on low-density ICs for consumers, remote controls, and smart cards. The system-level design talent and software expertise needed to compete in more sophisticated, higher-margin markets does not yet exist.
Lack of management and leadership. To train the next generation of design, system, and software engineers, China needs to attract or develop experienced management teams. Many expect Chinese semiconductor executives to repatriate as the Taiwanese did in the 1980s. But there are still cultural and systematic barriers, making this occur more slowly than it did in Taiwan.
Increased design cost and complexity. With the cost of chip design increasing exponentially, it is very difficult for a Chinese startup to focus on complex, leading-edge designs. One reason why Chinese companies are still focused on the low-end consumer market is because of talent, which is inextricably linked to the design complexity of the next-generation system chips. There have been several examples of Chinese design groups attempting and failing to design complex products such as microprocessors or DSPs.
Time-to-market pressures. Design complexities and the need to speed products to market have led to the emergence of the "design foundry," particularly important to small and medium fabless companies to bridge the gap between the design industry and the foundries. These professional design service companies help to improve frontend design capacity, accumulate and enrich the IP libraries, project risk analysis, and develop cooperation with suppliers. I met with several suppliers during my recent visit to China and Taiwan such as Faraday, Global Unichip, IP Core, VeriSilicon, Socle, and Goyatek. This is also an important approach for US companies, and companies like eSilicon are gaining momentum using this model.
High SoC development costs. The system-on-chip (SoC) dream has proven very difficult and expensive, and has created delays in time-to-market. But on the upside, alternative solutions are arising, such as system-in-package (SiP), which allows multiple chips (such as memory and ASICs) to be stacked together to meet the requirements of an emerging application solution. These solutions are quicker and perhaps more cost-effective than true SoCs. Major packaging houses are aggressively providing related technologies to the design community. This could be an attractive approach for smaller fabless firms to gain a competitive edge.
Less mature VC community and absence of IPO market. Another challenge for fabless companies in China is the lack of maturity in capital markets compared to the US and Taiwan. Even if a fabless company is successful in China, it must go to Hong Kong or the US to raise public money. Solving these challenges is key to attracting and maintaining desirable talent and executives, and will be part of the overall solution to moving toward more complex chips and SoC designs.
Success for the semiconductor industry will come from identifying or creating new applications. Successful companies must become global enterprises, which means collaboration with partners and peers to take advantage of opportunities and to ensure a cohesive design and supply chain. There is a real opportunity for non-Chinese companies to collaborate with teams in China, putting together a winning combination of architectural and system design expertise and customer and manufacturing proximity. Chinese fabless companies have already shown interest in meeting potential US and Taiwan partners. Many collaborative models will likely emerge to reflect a diverse market. With China's emergence as a powerhouse in this industry, the global nature of our fabless society is inevitable.
Contact Jodi Shelton, FSA, ph 972/866-7579, e-mail [email protected].