Since the 1990s, the semiconductor industry has experienced further structural change.
The radical changes in the methodology of IC design referred to as the
“system-on-a-chip”(SoC),which combines“system-levelintegration”on a chip with
“modulardesign”and “design automation”,havefacilitated thereuseofdesign building blocks (SIPs) to increase the level of design performance (Ernst, 2005a). In turn, this change has generated drastic changes in the organization of the value chain.
In a nutshell, the IC design industry has been transformed, as Ernst (2005a: 2) observes, from an integrated form of design organization, where (almost) entire ICs
are designed within a single firm, to a form in which the stages of chip design can be outsourced to other firms and relocated across national boundaries. This vertical specialization of IC design has transformed an artisan-type design activity into a highly routinized factory-type operation where most of the knowledge is codified and most of the stages of the design can be automated (ibid). The reorganization of the value chain of the semiconductor industry thus makes the IC design a globalized industry.Currently,designing an IC now can beorganized on a24 hours’basis,with teams relaying around the globe. It is under the above transformation that Taiwan has begun to face a new challenge. While Taiwan has emerged as a power house in the semiconductor industry since the 1990s, it faces a new challenge arising from
emerging global design networks through which global MNCs outsource their design functions to China and India to take advantage of the reduced labor costs. Taiwanese IC design firms in HSIP are thus sandwiched between innovative firms that are seeking low production costs and catching up firms that are increasing their technological capabilities. The sustainability of the HSIP thus deserves to be investigated.
3.2, The emergence of semiconductor cluster and its transformation in GPNs The emergence of the semiconductor industry in Taiwan was almost a state creation, which has been described in detail by Mathews and Cho (2000), Amsden and Chu (2003), Breznitz, (2005) in English and by Chen (2003) and many others in Chinese. In a nutshell, beginning in the late 1970s, the state decided to establish the semiconductor industry as a tool to deepen its industrialization and to upgrade its technological level. Rather than to depend on MNCs for transferring technology, the state established experimental factories through technology acquisition from abroad (mainly RCA), and then later spun off to the private sector (late became UMC).
During the process, the ITRI was the main actor and later became a facilitator in accessing new technology and then transferred to local firms. The Taiwanese semiconductor firms then worked closely with MNCs and gradually assimilated technologies and accumulated knowledge in due course.
The most significantcontribution ofTaiwan’ssemiconductorindustry wasthe establishment of TSMC, which was the first semiconductor foundry in the world.
Foundry companies in the semiconductor industry do not design chips, but only manufacture the chips designed by other companies. The establishment of TSMC became a catalyst that enabled many domestic fabless IC design houses to emerge and take advantage of existing fabrication facilities. The emergence of a vast number of small IC design houses led Taiwan to become one of the major IC design countries in theworld thateventually caused Taiwan’ssemiconductorindustry to concentrateon
the area of application-specific integrated circuits (ASIC) that could be used in various areas of the PC system. This in turn largely enhanced the competitiveness of Taiwan’sPC industry.Now Taiwan has 261 IC design firms whose sale value lumping together reaches over U.S. $10 billion and has about 23% of world market share. This scale of fabless IC design industry is second only to the U.S. (MIC, 2007). These IC design firms are clustered in HSIP. The related semiconductor firms, from IC design to foundry and Mask are densely located in the area that takes less than 30 minutes by car or by motorbike.
Table 3: Top 10 IC Design Firms in Taiwan 2005-2006
Firm
Media Tek 1 529.43 1 464.91 13.9 Digital Consumer DVDs, Digital TVs, PC Optical Storage Drives, Wireless Handsets IC Novatek 2 314.28 2 259.84 21.0 Flat Panel Display
Driver IC, Video Display Controller, Digital Imaging, Digital Audio/Video
Multimedia IC
HiMAX 3 246.23 5 177.60 38.6 Display Driver, Timing Controller, Television Chipsets
VIA 4 214.41 3 191.34 12.1 Chipsets, Processors, Mainboards, S3 Graphics, Vinyl Audio, Video Display IC Sunplus 5 170.76 4 187.81 -9.1 Digital TV, DVD,
Set-Top-Box (STB), Digital Photo Frame (DPF)
Phison 6 124.52 9 63.08 97.4 Flash Controller, Embedded Solution,
OEM/ODM Solution, Multimedia IC Realtek 7 124.23 7 106.36 16.8 Communications
network, Computer peripheral, Multimedia, Ultra-Wideband IC Etron 8 104.81 8 67.05 56.3 Memory IC, Portable
DVD, Car TV, PMP LCD Display Controllers
SiS 9 79.08 6 115.34 -31.4 Desktop PC Chipsets, Mobile PC Chipsets, Network, Embedded Solution, Components Mstar 10 72.60 10 56.04 29.6 Display IC Products,
Mixed-Mode ASIC/IP
Source:MIC, 2007.
3.3 the GPNs and local networking in HSIP
The development of IC design industry in HSIP in a large degree has to do with the strong production capability of Taiwanese IT industry, especially PC and handheld producers (Dedrick & Kraemer, 1998; Wang, 2007). Currently, Taiwanese PC firms produce over 90% of notebook PC for the world market (MIC, 2007) and many other components. These firms now play as major key suppliers to the branding firms that dominate the turnkey position in the GPNs. Because the severe competition among branding firms, these turnkey suppliers have to deliver the goods and services to the branding firms in a timely basis to fulfill the time-to-market demands. Low cost, speed and flexibility thus become the major competitiveness of the turnkey suppliers.
Logically, these PC firms in turn transmitted the pressures upon the IC designer for good quality and low price products in a timely basis. In order to achieve this, the PC system manufacturers have to work closely with IC designers in its initial stage of a new model when necessary in order to avoid mistakes and save time. Although all the major system manufactures have moved their production bases to China (Wang and Lee, 2007), the headquarters are still located in Taiwan (mainly in HSIP or nearby area) whose design teams are working closely with Taiwanese IC designers that
provide most of the chipsto theformers’end products.ThePC and handheld system firms thus have built a strong networking relationship that sustain the competitiveness ofTaiwan’sIT industry in theGPNs.
In addition to the IT system firms, IC design firms in HSIP also closely connect with foundries that constitute an essential element as a cluster. It is imperative for the IC designers to collaborate closely with engineers of the foundry in each stage of the chip design to avoid the possible low yield rate. Moreover, since the TSMC is the leading foundry in the semiconductor industry whose process technology is in the leading edge of the industry. Therefore, the spatial proximity of IC design houses and foundry not only largely reduces the transportation and transaction cost, but also benefits the IC design firms in learning new knowledge and technology. It is due to this technological reason that IC chip designers in HSIP have deeply networked with the nearby foundries such as TSMC and UMC. According to a survey (Deng, 2005), Taiwanese IC design firms use as high as 85% of local foundry service for their own products. This shows that spatial proximity matters for the IC design industry and this in turn create cluster effect.
3.4 Collective learning
The semiconductor cluster in HSIP has created an environment that facilitates collectivelearning and continuing upgrading.Thissustain HSIP’scompetitivenssin the semiconductor industry. First of all, the state continue to play an important role in facilitating learning. For example, in the 1990s, ITRI developed IC design software and standardsaswellassetup ‘Common Design Center’to assisttheIC design industry to diffuse the knowledge and to develop sophisticated IC components (Chang and Tsai, 2002:109). The ITRI also set up a SoC development Center, collaborating with MIT and UCLA, in developing frontier technology for IC design. It facilitates the formation of R&D consortium in developing new 3D IC and memory chips for next generations. The members of these consortiums share the knowledge and new technologies being developed, which largely help the IC design firms to upgrade their technology capability.
In addition, the collaborations among the IT system firms, IC design vendors, and foundries, and most of all the dense linkages between HSIP and Silicon Valley (Saxenian and Hsu, 2001) are also channels that facilitate collective learning. The IC designers are constantly engaged with IT system producers for design a new PC product even at the initial stage of the ODM. The IC designers also constantly consult with engineers from foundries due to the necessity of following the design rules of foundries’processtechnology.Theengineersin HSIP areflying back and forth between HSIP and Silicon Valley for business and technological exchanges. All these arebeneficialforHSIP’scollectivelearning asawholeand link thelocalto the innovation center in the world. Moreover, the adjacent Tsinghua and ChiaoTong
serve as the bridge that disseminating knowledge as well as forming alliance for co-developing new technologies. The above networking and bridging together
construct HSIP as a learning regional region of the semiconductor industry as a whole and IC design sector in particular.
4,Zhongguancun’sinnovation pattern and weaknesses
Beijng’sZhongguancun (ZGC)isarguably described asthemostinnovative region in China.Thereare68 universities(including China’smostprestigious
universities, Peking and Tsinghua), and 213 state-sponsored R&D institutes (including the Chinese Academic of Science, CAS) and over 300 thousand students in Beijing.
Moreover, Beijing hosts over 36% of the honorary fellows of the CAS and Chinese Academic of Engineering. All these indicate that Beijing has more affluent S&T personnel as compared to other cities in China.
During the early stage of economic reform, there were scholars who had utilized their scientific research results to build their own enterprises and created many famous mingying (none-governmental owned) high-tech enterprises such as Legend (spun-off from the CAS). In 1988 the central government decided to develop this area asthe‘Silicon Valley ofChina”.Sincethen,thelocalstatemobilizesresources to develop this area and to use policy measurements to attract firms to locate. Over the year, the development of the ZGC has created an agglomeration effect for the
high-tech, especially the IT industry. It gathered over 13000 firms in the area in 2006 (the Haidian area), including Legend, Stone, Fangzheng, and MNCs such as Lucent, HP, Ericsson, Hitachi, Siemens, etc (2008, Annual report of ZGC). But different from other two regions (Shanghai and Shenzhen) where the IT hardware is the dominant sector, the ZGC has concentrated more on the industries of IT software development and production than on those of the hardware assembly. Many MNCs also established their R&D centers in this area. Until 2006, 95 among the top 500 largest firms in the world have set up branches in this area, among them, 65 were R&D centers (ZGC report, 2008). Together with the high concentration of R&D personnel and institutes in Beijing area, ZGC has become the most important center for technology learning and innovation. Even the biggest domestic firm, such as Leveno, has established it R&D center in this area and moved it hardware production and assembly into Suzhou and Shenzhen areas. Currently, many of China’smostnotableICT companiescan be found in this area. Besides Lenovo (originally called Legend), they include Baidu (百 度),China’sleading Internetsearch enginecompany;UFIDA (用友),China’slargest privately owned software company; Datang (大唐),oneofChina’slargest
telecommunication solution companies; Aigo (愛國者), China’sleading portable storage and digital entertainment product maker, etc. The importance of IT industry in
ZGC can be shown in fig. 1.
4.1 GPNs and IC design sector
As discussed above, the global transformation of the semiconductor industry has led to the emergence of global innovation networks in which the IC design can be disintegrated and reintegrated in a global scale. The vertical specialization of IC design has transformed from an artisan-type of design activity into a highly routinized factory-type operation where most of the knowledge are codified and most of the stages of design can be automated (Ernst, 2005a). Now all parts of the IC design, including the whole procedure from specification to finished chips, can be outsourced to mainly either China and India where salaries of engineers are only 10% to 20% that of Silicon Valley. This saves large amount of cost for the firms. Thus, the IC design industry now can beorganized in a24 hours’based working team around theglobe (Brown and Linden, 2005:16).
It is under this global transformation of IC design sector that ZGC has been integrated into the GPNs of the semiconductor industry. Therefore, many IC design center of the ICT industry has set up their R&D center in ZGC to take the advantage of low costand from thelocalstate’sfavorabletax reduction.According to ourreview, most of the IC design works that the local engineers do are procedural and tedious parts rather than the architectural framework. The frameworks are decided in the center from abroad. What the local engineers do are the most sophisticated parts which need patient and longer working hours to finish. Since the MNCs pay good salary61, therefore they can recruit the most talented engineers at the expense of local IC design firms. In general, the R&D centers of MNCs in ZGC rarely have local
61 The salary for the engineers work for MNCs is average RMB 7000 per month, higher than that
connection with local firms. The only connection that these R&D centers have established is the elite universities where they have donated labs so as to recruit talented students in the future.
On the other side, the local Chinese IC design firms are still very small and at their earlier stage that are still far lag behind the Taiwanese firms in terms of technological level (table4). They do not have the similar milieu that the Taiwanese counterparts have enjoyed in terms of networking with IT system firms and foundries.
The biggest IC design firm, CEC Huada, is a state-owned IC design firms whose major products are like smart chip that does not contain very sophisticated technology.
According to our interview, the major difficulties that the IC design firms face are (1) lacking routine customer, because most of the IT system firms will not use their chips, they can only design chips that are used by mainly state sectors; (2) lacking of
incentives to attract talented people such as stock option, therefore they are not able to compete with MNCs; (3) lacking of networked foundries nearby, they therefore have to send their chips to be manufactured in Shanghai which is not favorable for chip designers to upgrade their knowledge level. All these factors stated above are not favorable for IC design firms in ZGC to survive in the market.
The only common resources which the IC design firms in ZGC has enjoyed and shared with the Taiwanese counterparts are the overseas returnees. Indeed, ZGC has attracted large amount of overseas returnees to start their own adventures which has increased rapidly over the years (ZGC annual report, 2008). The IC design firms which are setup by overseas returnees and are well connected with the GPNs are belonged to this type. For example, the rising star Vimicro (中星微) is founded by a returnee, whose specialties are on wireless communication chips and handheld equipments’paneldrivers.TheproductsofVimicro now areused widely in PCs and handsets. This type of design firms is the national hero that is highly supported by the Chinesestateforits‘independentinnovation’(z-zhu-chuang-xin) effort. Therefore the Chinese state supports this emerging firm to link closely with local handheld and service providers as to largely expand their market share. In this sense, this type of firm can establish networks with the local firms and enjoy the fruit of the booming economy. Nevertheless, whether this type of firm can sustain its innovation
momentum depends to large degree on the institutional change towards innovation milieu. On this, the current milieu of ZGC is still far from the conditions that can sustain a innovation cluster. Table 4 shows the top 10 IC design firms in China, in which 4 out of 10 located in ZGC. However, only the Vimicro is the private-owned enterprises, other 3 are state owned electronics firms.
Table 4: Top 10 IC Design Firms in China 2006-2007
Firm
1 14.61 2 12.00 21.75 Smart Card Chips, WLAN, Zeni EDA
2 12.90 5 9.04 42.70 Chipset, IP
surveillance solution,
3 11.06 10 3.32 233.13 Wireless
Communications,
4 10.79 4 9.19 17.41 Smart Cards, Communications
5 8.78 1 13.46 -34.77 CMOS mix signal SOC, CPU IP core, MP3, USB, SPDIF,
6 8.50 6 8.43 0.83 Audio/video
processing IC, MCU, Mixed-signal IC, Power management IC and power IC Hangzhou Silan
Ltd.( Hangzhou) Driver, DC Motor
8 7.06 3 10.13 -30.31 PC Multimedia Processors, Mobile
9 6.83 8 6.57 3.96 Smart Card Chips, Power meters, Home Source: 1. The Industrial Map of China Information Technology 2007-2008; 2. ICT Country Report 2008
4.2 The IT software industry
The IT software industry currently is the most innovative sector in ZGC. As in the IC design sector, this sector can also be divided into 3 types: MNCs, state-owned and private-owned, in which the private-owned is the most innovative type.
MNCs’strategy in softwareindustry in ZGC hasmainly focused on office application and middleware software. Due to the widespread and extensive piracy rate in China’ssoftwaremarket,theMNCshavealmostlost their competitiveness in the consumer software market field. Therefore, in this sector of the software industry, MNCs are the major system and application suppliers, including office application, enterprise management software such as enterprise resource planning (ERP), data base, etc. It is also in this sector that MNCs have interest to collaborate with Chinese firms in the expectation of high market growth.
Different from IT hardware, whose production procedures can be universalized, the IT software has to consider and adjust to the language, needs and customs of local users. That is, the utilization of office and middleware software has to be
indigenized. To fulfill this demand, the MNCs can only depends on local people and
local firms to expand the market share. Therefore, all the major software MNCs, such as Oracle, Sun, Cisco, collaborate with local firms to sell their products. It is also through this collaboration, the MNCs have trained local engineer related knowledge and through which knowledge diffusion has occurred. Zhou and Tong (2003) have documented the relationship between MNCs and local firms in ZGC. In brief, they argue that there are various types of relationship between MNCs and local firms. The first is a parallel relationship in which large state-owned software firms develop their
local firms to expand the market share. Therefore, all the major software MNCs, such as Oracle, Sun, Cisco, collaborate with local firms to sell their products. It is also through this collaboration, the MNCs have trained local engineer related knowledge and through which knowledge diffusion has occurred. Zhou and Tong (2003) have documented the relationship between MNCs and local firms in ZGC. In brief, they argue that there are various types of relationship between MNCs and local firms. The first is a parallel relationship in which large state-owned software firms develop their