by FAN Yicheng

When Infinigence AI moved into Shanghai's Xuhui-based Foundation Model Innovation Center, a government-backed AI cluster, the artificial intelligence chip start-up employed just seven people. Two years later, its workforce has grown to more than 200, and the company has raised nearly 500 million yuan in an A+ funding round.

Co-founder and chief executive XIA Lixue said the company's growth owed less to individual subsidies than to proximity. Access to computing power, suppliers and engineers in one district shortened development cycles and cut coordination costs, allowing the firm to scale while staying focused on product development.

Over the past five years, the city has shifted its artificial intelligence strategy away from experimentation toward commercial deployment. More than 10,000 AI-related companies now operate in Shanghai, spanning chips, algorithms, foundation models and downstream applications.

The Shanghai Municipal Commission of Economy and Informatization said this week that output from the city's AI industry is expected to exceed 550 billion yuan this year, with growth of over 30%.

From incentives to coordination

SHEN Hao, deputy chief engineer at the Shanghai Artificial Intelligence Research Institute, said Shanghai had shifted away from one-off incentives and was instead working to ease bottlenecks in regulation, funding, infrastructure and talent.

Shanghai began laying the groundwork in 2017, issuing its first city-level guidance on next-generation artificial intelligence and later setting up a cross-department coordination mechanism. This approach was formalized in late 2021 under the city's 14th Five-Year Plan (2021-2025), which placed greater emphasis on applications, commercial use and links with manufacturing rather than research targets alone.

A clearer regulatory environment followed in October 2022, when Shanghai introduced a dedicated regulation for the AI sector, spelling out rules around algorithm trading and data use and reducing legal uncertainty for companies and investors.

As large language models moved to the center of the industry, the city followed up in November 2023 with measures focused on data availability, model training and talent recruitment.

Rising costs drive AI clustering

By 2025, operating costs had become a more visible constraint. In July, Shanghai rolled out new measures under its "AI+" initiative to lower expenses tied to computing power, model usage and data access.

The city plans to issue 1 billion yuan worth of vouchers covering these areas, with subsidies offsetting 30% to 100% of costs. The aim is to ease the financial burden as companies move from pilot projects to large-scale deployment. Further measures released a month later targeted AI-enabled manufacturing and embodied intelligence, with a goal of 50 billion yuan in core industry output by 2027.

Policy tools have been paired with physical clustering along the Huangpu River. On the west bank, Xuhui's Foundation Model Innovation Center concentrates on large-model development and applications. On the east bank, Pudong's Model Power Community focuses on chips, hardware platforms, embodied intelligence and system integration.

Foundation Model Innovation Center, launched in August 2023, was China's first innovation cluster built specifically around large AI models. It has attracted more than 500 companies, including MiniMax and StepFun.

The Model Power Community, which opened in August 2024, brings together hardware development, model training, dataset construction and testing facilities, shortening the path from research to deployment.

In September 2025, Shanghai added Zhangjiang AI Innovation Town, linking several AI parks into a larger zone. The project aims to attract 1,000 additional companies within three years and lift industry output beyond 250 billion yuan, supported by incentives such as free office space, talent housing and computing vouchers.

Talent as the binding constraint

Despite the infrastructure build-out, talent remains the most difficult constraint.

Shanghai has responded with institutional experiments such as the Shanghai Innovation Institute jointly established by the Ministry of Education and the municipal government.

Founded in September 2024, the institute operates outside the traditional university system. Embedded in an industrial cluster, it enrolls only doctoral students and uses a dual-mentor structure to link research more directly with industry needs. By September 2025, it had attracted more than 3,300 applicants and selected over 800 students.

Shared computing platforms and open-source frameworks are intended to support this pipeline, while international events such as the World Artificial Intelligence Conference connect local firms with global researchers and investors.

For firms like Infinigence AI, the impact has been practical. Operating within a dense network of computing resources, partners and policy support allows firms to focus on research and product development while scaling inside the city, strengthening Shanghai's appeal as a place to grow.