The GBA, located in the coastal areas of southeast China, is the frontier of venture capital in China. The GBA includes nine cities in the PRD, namely Guangzhou, Shenzhen, Dongguan, Foshan, Zhongshan, Zhuhai, Huizhou, Jiangmen, and Zhaoqing. It also includes two special administrative regions, namely Hong Kong and Macao, with a total area of 56,000 km² (Figure 1). In 2020, its GDP reached RMB 11,365,477 million yuan, accounting for 11.2% of China’s total GDP, and its resident population reached 8,617,200, accounting for 6% of China’s total population¹(1 Economic statistics were obtained from the National Economic and Social Development Statistical Bulletin; Hongkong and Macau data was obtained from the Statistics Department of the Special Administrative Region Government; population data was obtained from the Seventh Census Bulletin).

Owing to the convergence of diversified capital and multiple governance forces, and as one of the most typical export-oriented economic zones in China, the development of the GBA has been of interest to the academic society. Hong Kong and Macao have gone through a long integration process with the PRD, from the “front shop, back factory” labor division to comprehensive integration, and because of their special cross-system characteristics, the governance structure has been conceptualized as that of typical hybrid capitalism (Ye et al., 2019; Wu et al., 2021a). In 2019, the Outline of Development Plan for Guangdong-Hong Kong-Macao Greater Bay Area further clarified the GBA as a national strategic area. In this plan, the central government clearly proposes to build the GBA into a vibrant world-class urban agglomeration as well as an international science and technology innovation center with global influence. Such positioning reflects the development focus and ambition of the GBA at the next stage. This was immediately followed by the Guangdong Province and launched the Three-year Action Plan for Guangdong Province to Promote the Construction of Guangdong-Hong Kong-Macao Greater Bay Area (2018-2020). This plan establishes 100 specific planning and implementation measures in detail for the next three years. According to the plan, building an international financial hub with Hong Kong and Macao is defined as the dominant task, specifically by supporting Guangzhou to build a regional private equity trading market, promoting Shenzhen to develop a capital market with the Shenzhen Stock Exchange as the core, and promoting Macao-Zhuhai cross-border deep financial cooperation. Promoting the establishment of a venture capital market to support the next stage of “innovation-driven” development and the transformation of the GBA will be one of the priorities of the GBA in the future.

The venture capital data used in this article was sourced from the China Venture Capital Source (www.cvsource.com.cn). The data cover the venture capital events in China from 2000 to 2018 with firm-level data. The data attributes include the name and address of the venture capital institution, name and address of the venture capital receiving enterprises (startups), industry type, investment amount, capital ownership structure of the investment institution, etc. In this study, some data such as detailed investment information and investment institution information were supplemented by consulting “Qichacha (https://www.Qcc. com),” “Tianyancha (https://www.TianYancha.com),” and “ITjuzi (https://www.itjuzi.com/)” and searching for news related to investment events on the Internet. Investment events with missing information that could not be supplemented were excluded from the study. Finally, 29,816 data on investment events on the national scale and 2,260 data on investment events at the GBA scale (both the investment institutions and startups were in the GBA during the investment event) were retained. Based on the information of the cities where the investing companies and the invested companies are located, we transform the venture capital between enterprises (with direction) into the venture capital between cities. Furthermore, based on geocoding technology and through the Baidu API interface, the spatial databases of venture capital institutions and startups in the GBA were established, with 5406 and 3167, respectively. Notably, the investment database in this study does not include data from Macao, because it includes gambling and tourism as its pillar industries and does not have a high degree of involvement in the venture capital network in the GBA.

Geocoding technology was used to establish the spatial databases of startups and venture capital institutions in the GBA, and the spatial patterns were observed by grid processing (8 km*8 km) of the GBA (Wu et al., 2021a). The result shows that startups are concentrated in cities in the core circle of the GBA (Figure 3), forming high-value areas in Guangzhou and Shenzhen. Venture capital institutions are significantly more concentrated in a few areas, mainly distributed in the core areas of Guangzhou, Shenzhen, and Hong Kong. The differences in aggregation scale between the two were compared by the DO approach (Figure 4). Results show that the agglomeration scale of venture capital institutions is smaller, specifically characterized by startups forming a distance distribution peak at a scale of about 5 km, whereas the agglomeration peak of venture capital institutions is significantly smaller. Therefore, venture capital institutions are more concentrated.

The inter-city capital connection exists not only between cities on the GBA scale. There must be cross-regional capital connections since Hong Kong and Shenzhen are both typical world-class financial cities. It might ignore the capital hub function of the GBA on the national scale by only analyzing the capital network within the region (Li et al., 2017). On the national scale, this article screens the cities closely related to capital in the GBA and integrated several key indicators, such as the number of startups (invested enterprises), number of venture capital institutions, number of venture capital events, and the total amount of venture capital. This article takes the selected top 20 cities in China as active cities in the field of venture capital. The total number of investment events in these 20 cities, together with the GBA, accounts for 83% of the total data, which is enough to support the cross-scale analysis of venture capital structure. Furthermore, considering the time of the establishment of the Growth Enterprises Market (2009) as a node, venture capital showed two different growth trends around 2009. This paper built this cross-scale urban capital collaborative network by diving into time periods to further study the evolution of the urban network. Owing to the limitation of the spatialization diagram, we only display the connections between cities with more than 10 mutual investment events instead of all the connections between cities.

Figure 8 shows the cross-scale network structure of venture capital connections in the GBA. The sizes of city nodes shown in the figure represent the total amount of venture capital in the city (invested + received, representing the city’s capital activity), and the capital connections between cities are expressed by the number of events of mutual investment between cities. The following conclusions can be drawn from the network structure in Figure 8.

First, on the regional scale, the GBA has formed a network structure of venture capital cities dominated by three core cities, namely Guangzhou, Shenzhen, and Hong Kong, and the periphery of the network structure mainly includes cities such as Dongguan, Zhuhai, Zhongshan, and Foshan. After 2009, the total amount of venture capital in the remaining GBA cities only accounted for 8.74% of the three core cities. The strength of capital’s external connection is only 26.27% of that of the three core cities. On the scale of the GBA, big cities tend to be the core of the capital network structure, suggesting that innovations in big cities can obtain capital support more conveniently. Although there are institutional obstacles, the geographical and cultural proximity of Hong Kong to Shenzhen and Guangzhou have also contributed to the formation of intensive capital links between these three cities.

Second, on the national scale, the GBA has formed a more intensive capital connection network with major core cities in China. Beijing, Shanghai, and Hangzhou, together with Guangzhou, Shenzhen, and Hong Kong within the GBA, have formed six national-level capital cities, which are prominent on the national scale. In addition, Nanjing, Tianjin, Suzhou, Chengdu, and other regional capital cities have also formed intensive capital connections with the GBA. Other cities in the GBA have not become capital cores on the regional scale, which indicates that the current venture capital connections still mainly occur between core cities. The national-scale analysis reflects the impact of capital hub cities on the structure of venture capital networks. Although intensive capital linkages have been formed between cities in the GBA, more intensive capital linkages still occur between the GBA and Beijing, Shanghai and Hangzhou. As the city with the highest degree of external ties in the GBA, Shenzhen’s capital ties with Guangzhou and Hong Kong before 2009 were only 14.36% of those with Beijing and Shanghai. Although this figure has risen to 16.5% since 2009, it still reflects the impact of national capital hubs such as Beijing and Shanghai on the GBA capital network. This also verifies the importance of cross-scale analysis for venture capital network research. The results of this study elucidate that in the venture capital network, the influence of factors such as institutional proximity, cognitive proximity, and organizational proximity is more important than geographical proximity. In addition to the capital hub cities of the GBA, other cities are also obviously affected by the national scale. Foshan and Dongguan are both important manufacturing bases of the GBA. Since 2009, the capital interaction between Foshan, Beijing, and Shanghai accounted for 73.68% of that with Guangzhou and Shenzhen, while the value of Dongguan has reached 150%. The results show that the regional scale network was shaped by the global scale network.

Third, from the perspective of the evolution of the urban network, only Beijing and Shanghai were national-level capital cities on the national scale, which dominated the capital network structure on the national scale before 2009. From 2009 to 2018, the GBA was fully integrated into the national venture capital connection network. Guangzhou, Shenzhen, and Hong Kong have quickly emerged as the national-level capital cores. Similarly, before 2009, a capital network structure involving only a few cities was formed on the national scale, and regional cities such as Tianjin, Wuhan, and Hangzhou participated in the capital network in the GBA. From 2009 to 2018, with the tide of innovation and entrepreneurship as well as the springing up of venture capital, more regional capital core cities and capital nodes appeared, and the whole network structure became more complex. This also shows that further improvement of the national capital market system as well as innovation and entrepreneurship system promoted the continuous improvement of capital collaboration between cities.

Venture capital behavior has the significant characteristic of “proximity” (Tian et al., 2020). The significant characteristic of high risk means that the selection of investment objects and the protection of investment returns have become core concerns in venture capital behavior (Wray, 2012). Therefore, “pre-assessment” and “post-monitoring” that can narrow the industry information gap have become the core operation mechanism of venture capital (Masucci et al., 2021). One of the strategies to reduce investment risk and improve communication outcomes is to select the investment target based on the criterion of “proximity” (Zook, 2002; Pan et al., 2019). This “proximity” includes not only geographical proximity but also multi-dimensional proximity conceptualized in evolutionary economic geography, including institutional proximity, organizational proximity, cognitive proximity, etc. (Boschma, 2005). Through “proximity”-based investment, venture capitalists can utilize their own social information network to obtain information on the investees and evaluate the transactions, and at the same time provide operational assistance to the invested enterprises based on the industry knowledge accumulation and their own network relationship (Enkel et al., 2020). The mechanism of multi-dimensional proximity explains why there is a clear pattern of collaboration between the core and the periphery in the GBA. At the same time, Guangzhou, Shenzhen, and Hong Kong have formed closer cross-scale capital connections with Beijing, Shanghai, and Hangzhou on the national scale, respectively. China’s rapid development of high-speed railways in recent years has weakened the geographic distance between the GBA’s interior and exterior. The compression of space-time distance promotes the long-distance spread of capital. The promoting effect of space-time compression on the growth of venture capital has also been confirmed by many empirical studies (Dong et al., 2020; Duan et al., 2020; Tian et al., 2020). In the formation mechanism of this mode, while geographical proximity has played a great role and geographical proximity has driven cities such as Dongguan and Foshan to obtain a large amount of manufacturing investment from Hong Kong and Shenzhen. Dimensions such as institutional proximity and cognitive proximity have also played a tremendous role, forming closer capital connections between the major capital centers in the country.

With the support of the central government, Shenzhen successfully developed its capital market after the reform and opening up. Shenzhen Stock Exchange was established in 1990 (Zacharias et al., 2010). In 2009, the Growth Enterprise Market was established to provide a platform for the growth of technological innovation enterprises. In 1991, the Shenzhen Stock Exchange’s annual transaction value was RMB 3.56 billion yuan, which has grown to RMB 162.22 trillion yuan in 2020. The transaction value of the ChiNext market also reached 46.67 trillion yuan in 2020, 255 times that of 2009. In 2010, the State Council approved the Qianhai Master Plan, and one of the core objectives of Shenzhen’s urban development was to gather offshore RMB outside the Chinese mainland (Chen et al., 2015). Because of a series of special policies, major banking institutions and financial practitioners swarmed into Shenzhen, consolidating its position as an international financial center (Chen et al., 2015). In 2021, Shenzhen ranked fourth in the world in the number of IPOs, making it a financial center that cannot be ignored in the world. Hong Kong has always played a critical role in the overall financial system of the country because of its special history of urban development and the “one country, two systems” policy (Wang et al., 2021). Institutional factors have promoted these two cities to become the core of venture capital for the GBA.

As the national capital hub, the formation mechanism of the GBA also includes two key mechanisms, namely, urban economy and path dependence (Chen et al., 2015). As the unique international financial cities in the GBA, Hong Kong and Shenzhen have gathered the major financial enterprises and financial practitioners in China. This high concentration of financial factors has produced a great Marshallian-localization economy (Marshall, 1890; Florida et al., 2017; Wu et al., 2021b). Through the geocoding of venture capital entities, this study found that 62.47% of startups are located in Shenzhen and Hong Kong, and the proportion of venture capital institutions is 80.04%. This agglomeration effect, strengthened by path dependence (Boschma et al., 2006), has deepened the specialization effect of Hong Kong and Shenzhen on the national and even global scale and has continuously promoted technological spillovers and innovations in the financial service fields of the two places, thereby continuously improving the position of the financial center and development of the capital market. The Gini coefficient of venture capital distribution was as high as 0.8004 in 2000. Although it decreased to 0.7435 in 2018, it can still be seen that the capital distribution is extremely uneven, that is, 69.99% of the GBA venture capital flows are related to Shenzhen and Hong Kong. The degree centrality results based on capital flow also reflect the path-dependent nature of this development. Before 2009, the degree centrality of Hong Kong and Shenzhen were 25.885 and 14.442, respectively. After 2009, they were 12.958 and 26.195, much higher than other cities. These mechanisms have further promoted the formation of the current GBA’s complex venture capital network structure.

As urban agglomeration becomes an important spatial unit of capital accumulation and the innovation transformation of the economic development, as important capital support for innovation, the flow and dynamics of venture capital have become the frontier of research on urban agglomeration spatial networks. Although existing studies have explored venture capital-based urban network structure, there is still a lack of attention to industry heterogeneity and cross-scale research. In particular, the GBA has rarely been examined as a case study. Combining the DO approach and methods of social network analysis based on a long-term series of venture capital data, the venture capital network structure of the GBA was deeply explored, including its industry heterogeneity characteristics, cross-scale network structure, and evolutionary characteristics.

First, the distribution of venture capital institutions and startups in the GBA shows a remarkably high concentration, with distribution mainly in the core areas of Guangzhou, Shenzhen, and Hong Kong. The network structure of venture capital also presents a core-periphery structure mainly consisting of a few core cities. Additionally, the flow of venture capital shows significant structural differentiation, mainly concentrated in five industries, namely manufacturing, the Internet industry, the financial sector, the cultural media industry, and the medical and health industry. The flow of venture capital in these types of industries is closely related to its local industrial structure background.

Second, a cross-scale analysis of venture capital network structure reveals the hub function of the GBA on the national scale. On the regional scale, the GBA has formed a network structure with Guangzhou, Shenzhen, and Hong Kong as its absolute core. It has formed a dense venture capital connection network on the national scale. In addition, after nearly 20 years of development, a dense venture capital network has been established on a national scale, with Beijing, Shanghai, and Hangzhou, and Guangzhou, Shenzhen, and Hong Kong dominating the network structure.

Thirdly, three key mechanisms, namely, multi-dimensional proximity, institutional factors, urban economy and path dependence, have promoted the network structure of venture capital in the GBA. The “proximity” characteristic of venture capital behavior has promoted the GBA to become a national-scale venture capital hub. The institutional factor is an important reason that Shenzhen and Hong Kong have become the core cities of venture capital. Urban economy and path dependence have promoted the spillover and innovation of financial functions in the GBA core cities, thus, continuously promoting the growth of their financial center status and development of the capital market.

As an important element of innovation flow, the research on the structure of the venture capital network has important theoretical implications in both urban agglomeration network research and innovation network research. The network structure of urban agglomerations from the perspective of capital flow reveals the characteristics of cross-scale capital connections and the status of the GBA as a capital hub in China. In the study of network structure, we found a network structure that is different from traditional flow elements such as traffic flow and logistics, that is, the higher concentration of capital. The formation mechanism of this network structure combines innovation network theory with flow space theory. Urban geography reveals the role of diversity and urbanization economy on local innovation. These key mechanisms promote the agglomeration of capital. The relational turn of economic geography coincides with the perspective of flow space, such as local buzz and global pipelines, global production network reveals the characteristics of local links and cross-scale links of knowledge, and the research based on venture capital in this paper also responds to these theories.

Our research is one of the few papers that systematically reveals the industry structure differences, cross-scale connections, and evolutionary characteristics of venture capital. Based on the GBA case, Hong Kong is also included in the venture capital network, which forms a case supplement to previous research on venture capital. The analysis based on the spatial data of venture capital entities shows the differences in the distribution of different venture capital entities, and the analysis of the network structure based on the cross-scale urban agglomeration also shows the GBA’s status as a national venture capital hub and its growth.

Notably, this research has thrown up many questions in need of further investigation. First, the collection of venture capital data is difficult, coupled with the fact that venture capital events cannot disclose the exact investment amount owing to confidentiality and other reasons, which makes the research of the venture capital network have possible sample bias and cannot fully demonstrate its network structure. Various channels supplement our database, but the construction of the database still needs to be improved. Second, the flow of innovation capital is not limited to regional or national scales, and there is no political boundary to the flow of capital. The research based on the GBA and national scale cannot fully reveal the capital hub function of the GBA. Further in-depth analysis of venture capital data on a global scale is needed. It is believed that Hong Kong’s and Shenzhen’s capital hub functions will be further highlighted with more comprehensive data. Third, the formation mechanism of the GBA’s venture capital network structure has not yet been comprehensively and deeply analyzed. Further in-depth research is needed combining qualitative and quantitative methods to quantify key influencing factors, interviews with venture capitalists, startups, and competent government authorities are also needed for further investigation.