- Software revenue in Nanjing tops 1 trillion yuan in 2025, marking its first trillion-yuan industry cluster
- City leans into industrial AI and “agent” systems as it defines a different path from Hangzhou’s platform-led model
The software industry in Nanjing, Jiangsu’s provincial capital, generated 1.02 trillion yuan ($151 billion) in revenue in 2025, up 16.7% year-on-year, making it China’s fourth-largest software hub after Beijing, Shenzhen and Shanghai, according to local media reports.
The sector has also become the city’s first trillion-yuan industry cluster.
Building on that base, Nanjing has begun a coordinated push to develop “AI plus software,” positioning itself as a national leader in intelligent software systems and, more ambitiously, as a potential No.1 “city of agents,” a term increasingly used to describe AI software stack that can execute tasks autonomously.
A local entrepreneur put those ambitions into words. “Hangzhou is e-commerce, Shenzhen is hardware. Nanjing has every chance to become the capital of intelligent agents,” Wang Liang, vice president of Nanjing-based software firm Huizhi Intelligence (汇智智能), said at the 2026 Nanjing Software Conference on June 16.
Competing with Hangzhou
The comment captures the city’s intent to avoid competing directly with Hangzhou’s internet platform ecosystem, instead targeting the emerging AI agent layer.
Hangzhou, meanwhile, is advancing its own AI ambitions. The city has set a goal of becoming a national AI innovation hub by 2030, with core AI industry revenue expected to exceed 1 trillion yuan.
It already reports about 460 billion yuan in AI-related output and hosts major foundation model players including Alibaba’s Qwen and DeepSeek, alongside four national AI pilot bases.
The contrast reflects two different industrial starting points.
Hangzhou’s advantage lies in its internet platform economy and private capital depth, with venture capital funds estimated at roughly 2.2 times the scale of Nanjing’s and a private-sector share exceeding 90%.
Nanjing’s strength sits in industrial software and manufacturing integration. Its smart grid sector alone accounts for around 80% of China’s market, while its industrial software stack spans R&D, production control and enterprise management.
The city’s software ecosystem has long been underpinned by state-run industrial players such as NARI Group (南瑞集团) and Guodian Nanjing Automation (国电南自), operating on longer development cycles but producing system-level industrial upgrades when deployed.
Examples of emerging “agent” capabilities are already visible. Nanjing-based software startup Whale Cloud (浩鲸科技) has developed an AI-driven software factory covering the full R&D lifecycle, while HopeRun (润和软件), a leading software developer, is applying Huawei’s open-source HarmonyOS systems in public services and transport digitization.
Talent remains a decisive battleground. A 2025 Nature Index supplement ranked Nanjing fifth globally in research output, ahead of Hangzhou at 13th. The city has 13 so-called “double first-class” universities, but its technology commercialization rate, at 28% in 2025, lags far behind Hangzhou’s 45%.
Hangzhou’s talent pipeline is closely tied to Zhejiang University, which has produced founders or faculty-linked entrepreneurs behind several of the city’s prominent startups, including DeepSeek, Manycore Tech (群核科技) and DEEP Robotics (云深处科技).
Whale hunting vs. Shipbuilding
The divergence has also been described in cultural terms. Hangzhou’s innovation style is often framed as “whale hunting,” concentrating resources on a few high-potential bets such as foundation models and scaling them rapidly through platform capital and talent concentration. It is a variant of Silicon Valley’s “hits” strategy.
Nanjing, by contrast, is often described as practicing a “shipbuilding” approach—long-cycle, engineering-heavy and deeply embedded in industrial clients, exemplified by customized power-system software developed for China’s energy infrastructure.
While Hangzhou currently dominates the visible AI startup landscape, Nanjing is still working behind the scenes, quietly building up its capabilities away from public view.
It is pursuing a more system-oriented strategy, backed by a policy package issued in April.
The framework targets more than 100 AI software vendors and agent developers, 300 industry-specific agents, and wide adoption of AI coding tools across thousands of firms.
At a recent software industry event in Nanjing, companies including Huawei, AI chipmaker Cambricon (寒武纪) and AI infrastructure developer Sugon (中科曙光) signed agreements with local authorities to build a domestic AI software and hardware ecosystem, underscoring growing industrial coordination.
Beyond the zero-sum rivalry
Many view the Nanjing-Hangzhou race as yet another example of China’s relentless intercity competition, but it is not a zero-sum game on closer inspection.
Hangzhou leads in internet ecosystems and capital agility, with early advantages in foundation models and rapid commercialization.
Nanjing, by contrast, draws on deep manufacturing roots and industrial software strengths, giving it an edge in downstream, system-level AI deployment.
In simple terms, Hangzhou excels at “0 to 1” breakthrough innovation, while Nanjing is stronger in “1 to 100” scaling and integration. Over time, overlaps will emerge, but their underlying positioning remains distinct.
Linked by the Yangtze River AI Corridor — an emerging world-class AI industrial cluster along China’s Yangtze River Economic Belt, spanning from Shanghai in the east to Chengdu in the west — the two cities are charting different paths that together expand the range of China’s AI industrialization model.
Whether Nanjing can realize its No.1 “city of agents” ambition — and beat Hangzhou at its own game — depends on how quickly it can convert academic strength into industrial output.
Ultimately, the outcome may not be about which city wins, but the emergence of a broader ecosystem.
