To many, the meteoric rise of DeepSeek is a “Sputnik moment.” Despite its modest scale and resources, the Hangzhou-based AI startup has come a long way, surmounting tech barriers, export controls, and trade restrictions. Chinese media and tech insiders have hailed its emergence as a “historic moment” for China’s innovation model.
Much of the discussion thus far has centered on the company’s founders and team members, and on whether U.S. semiconductor restrictions have failed to curb China’s AI ambitions. However, these narratives generally miss a bigger question: who actually incubated DeepSeek and other Chinese AI startups?
We know that Beijing’s state-led industrial mobilization has played a pivotal role in shaping such transformative industries as solar power and electric vehicles. Through early-stage subsidies, tax incentives, and large-scale infrastructure projects – hallmarks of China’s industrial policies – Beijing catalyzed market growth and spurred technological innovation.
Yet, China’s tech startups have followed a distinct development path – one shaped by local governments, who have tried to build localized incubation ecosystems. High-quality innovation does not require many winners, and inefficiency needn’t mean failure: a single outlier, such as DeepSeek, can be enough to trigger an AI tsunami.
One pillar of local incubation is talent mobilization. With strong local universities and proactive talent recruitment policies, cities like Hangzhou provide a dynamic research environment and a robust local market for tech talents. DeepSeek’s core members all graduated from China’s top universities. The founder is a graduate of the best university in Zhejiang Province, where Hangzhou is the capital.
Local industrial policies are another pillar of local incubation. Industrial parks and “characteristic towns” (特色小镇), for instance, have provided handsome land and rental subsidies, tax incentives, and talent recruitment support for startups. As a result, startups have been flocking to these parks and towns. DeepSeek is located in a district housing tens of thousands of tech startups and multiple high-tech industrial parks.
A recent example is the development of AI Town as part of the Hangzhou West Science and Technology Corridor, a 33-kilometer sci-tech cluster that brings together leading firms, research institutes, government agencies, and AI startups. BrainCo, a world-renowned startup specializing in brain-computer interface technology, is located right inside AI Town.
Internet giants also aided local incubation. Alibaba has provided both financial support and technological integration platforms, fostering synergies between local AI startups and tech giants. As a result, Hangzhou has created a robust tech ecosystem spanning AI, robotics, and brain-machine interfaces. Firms such as Unitree Robotics, BrainCo, and Game Science form a local web of technological collaboration.
Hangzhou’s success has a deep policy foundation: Zhejiang took an early bet on digital transformation. As early as 2003, when now-President Xi Jinping was the provincial party secretary (2002–2007), he launched the “Eight-Eight Strategy,” emphasizing the centrality of digital technologies for further economic reforms. And when he was the provincial governor (2013–2016), Li Qiang (now China’s premier) spearheaded bold administrative reforms that significantly reduced local market entry barriers.
Hangzhou has since attracted high-tech firms from across China. In 2018, Game Science relocated part of its operations to Yichuang Town – a specialized industrial park adjacent to the local campus of the China Academy of Art. This has allowed the company to recruit an army of talented artists for developing blockbuster products such as “Black Myth: Wukong.” The company’s founder highlighted local policy support as a key reason for the relocation.
Local officials across China now want their own DeepSeek phenomena. In his recent speech on local high-quality development, the party chief of Guangdong Province mentioned “industrial policies” and related terms 79 times, spotlighted companies like DeepSeek and Unitree Robotics, and called on cities in his province to catch up.
But different localities have different comparative incubation advantages. For example, the city of Shenzhen in Guangdong hosts tech giants such as Huawei, Tencent, and BYD. The development model there has historically favored such industry giants, rather than fostering a broader innovation ecosystem.
Giant firms naturally attract and nurture downstream industries closely linked to their own supply chains. Shenzhen’s semiconductor and integrated circuit industries rely heavily on demand from major companies such as Huawei and BYD. This has led Shenzhen to allocate resources in favor of large corporations.
Hangzhou and Shenzhen thus exemplify two distinct models of local state-led industrial development. Shenzhen’s “efficiency is life” mentality has fueled local industry giants but has left little room for startups. Innovation there has been demand driven, with startups serving as suppliers or supporters of large firms, rather than driving fundamental technological breakthroughs.
Shenzhen has scarce land resources, and these are dominated by large firms. With a land area only one-eighth that of Hangzhou’s, Shenzhen has allocated the lion’s share of its industrial land to a select few. For instance, in the first local land auction of 2024, BYD secured a massive 1.04 million-square-meter industrial plot. This inevitably crowds out smaller firms.
Shenzhen also has a very limited local talent pool; it has no leading academic institutions, unlike Zhejiang, where top-tier universities serve as a rich reservoir. Hence, Shenzhen relies on external talent acquisition, but limited land resources have also translated into costly housing for prospective recruits.
In contrast, Hangzhou’s innovation ecosystem is mostly incubation driven. By aligning tech companies’ interests with local government goals, the city has effectively integrated resources to create an environment where startups can withstand failures while incubating the makings of disruptive breakthroughs.
While the Hangzhou model cannot be replicated, local governments elsewhere in China are nonetheless pushing for transformations of their own – from being mere directors of development to active incubators of innovation. They are rewriting the development playbook, nurturing startups, and cultivating new industries. For China’s future economy, agile governments at all levels are just as important as angel investors.