Skip to content

Thousands of companies are driving China's AI boom. A government register keeps track of them all

    When DeepSeek burst burst onto the world stage in January 2025, it seemed to appear out of nowhere. But the Great Language Model was just one of thousands of generative AI tools released in China since 2023 – and there is a public archive of each of them.

    The country's top internet regulator, the Cyberspace Administration of China (CAC), requires any company launching an AI tool with “public opinion properties or social mobilization capabilities” to first archive it in a public database: the Algorithm Registry. In an entry, developers must show how their products avoid 31 risk categories, from age and gender discrimination to psychological harm and “violating core socialist values.”

    Applicants submit their applications to their local CAC (e.g. the Shanghai CAC for Shanghai registered companies), which forwards applications to the central CAC for final approval. Only then will a tool be publicly included in the algorithm register. While the European Union is pursuing one comprehensive AI law, notes Matt Sheehan, a researcher at the Carnegie Endowment for International Peace, China's approach to regulation is more ad hoc, focused on specific algorithms and building iterative standards. (The US has no comparable registration system or centralized regulatory agency.)

    Over time, the CAC has inadvertently created the most detailed map of a country's AI ecosystem anywhere in the world.

    *Data current as of April 2025, includes both 'generative AI' and 'deep synthesis' algorithms

    Open the August 2024 CAC update and you'll find DeepSeek listed as item 152, a single row in a neatly packed table. Scroll through the table and you'll find an AI that manages homestays and an AI that drafts patents. One assists midwives in a maternity ward in Shanghai; another helps manage the state's power grids. Kendra Schaefer and her colleagues at Trivium China, a Beijing-based policy consultancy, have compiled the CAC's updates into a comprehensive database, enriched by their own research.

    A wide view of the tree

    Nearly 80 percent of China's generative AI registrations are clustered in and around its main technology hubs: Beijing, Shenzhen, Shanghai and Hangzhou. Each city has its strengths: Beijing's elite universities, national laboratories and political strength give the country an edge in large-scale innovation; Shenzhen (in Guangdong) is home to a dense hardware supply chain and a vast pool of engineering talent; Shanghai, close to multinationals, excels in commercialization; and Hangzhou (in Zhejiang) is powered by Alibaba's e-commerce empire.

    But innovation is spreading far beyond coastal areas. Chongqing is positioning itself as an AI manufacturing and logistics hub; and heavy state investment have made Hefei, Anhui province, known as “China's speech valley” for its cluster of speech recognition companies, including iFlyTek. The files also come from less obvious regions such as Guizhou, China's 'Big Data Valley', where massive data centers support Huawei's Pangu model, and Inner Mongolia, where state-owned companies are integrating AI into mining and agriculture.

    *Data current as of April 2025

    In the Trivium dataset, state-related listings – from state-owned enterprises to government-backed research institutes – make up 22 percent of registrations. Many state-owned companies are working with Big Tech to build their AI: for example, PetroChina worked with Huawei and iFlyTek to create oil and gas applications; State Grid used DeepSeek to build a model that optimizes electricity grids.

    Foreign companies represent only 0.5 percent of the files. For example, Ikea has a smart shopper algorithm that generates product recommendations. Yum China, the parent company that operates Kentucky Fried Chicken in China, has launched a model that generates menus and promotional materials.

    Zoom in on the competition

    *Data current as of April 2025

    More than half of the entries in the algorithm register concern what Schaefer calls cross-sector technologies. These range from basic models to 'general purpose' text generators to a wide range of multimedia tools: voice changers, 3D renderers and image makers. “Nobody wants to get caught in a situation where they have to rely on a competitor's technology,” Schaefer says. Unlike the US, where OpenAI, Anthropic and Google DeepMind dominate the market, Chinese competition to build foundational AI remains diverse and contentious. But building these models is expensive and the market is starting to consolidate. China's six 'AI tigers' – Moonshot, Minimax, Zhipu, Baichuan, 0.1AI and Stepfun – are all backed by Alibaba or Tencent. ByteDance's Doubao surpassed DeepSeek as China's most popular chatbot, but its place at the top is not assured.

    Niche natives

    While the giants battle for chatbot supremacy, startups are hard at work in every sector imaginable.

    Squirrel AI 松鼠

    Founder Derek Li says his 12-year-old company is a step ahead of its competitors in the ed-tech space. They're “putting wheels on a horse,” he says, and building AI on top of their existing, outdated software. Squirrel claims to diagnose knowledge gaps, measure progress and adapt lessons in real time.

    When China banned for-profit tutoring in 2021, the company's revenues collapsed overnight. It revolved around licensing its platform to franchisees who also sold the company's AI-powered tablets. Squirrel's network includes more than 3,000 centers across China, serving 1.2 million students. Now the company is looking to expand into the US.

    Li, who withdrew his sons from a private school in Shanghai so they could be homeschooled on Squirrel's platform, says that “teachers in the future will not teach knowledge.” Instead, he says, “they will become data analysts, who understand learning reports and students' skills, and psychologists, who understand emotions and shape their personalities.”

    AI Kanshe 看舌

    AI Kanshe (translated as “AI Sees Tongue”) is a traditional Chinese medicine startup that analyzes health through images of the tongue, palms and face. The company was founded by Li Wenhua, a former employee of Yaoshi Bang, one of China's first online pharmaceutical platforms. Li had been studying tongue and hand diagnostics for many years and wanted to combine the diagnostic methods of traditional Chinese medicine with modern machine vision. The company serves both consumers and healthcare providers in clinics, pharmacies and some hospitals, providing diagnostic and decision support tools. The model is trained on more than 100,000 annotated images of tongues, hands and faces.

    Zhongtan Puhui Cloud Technology 中碳普惠云科技

    Founded in 2024 by Wu Song, a former quantitative trader on Wall Street, Zhongtan Puhui Cloud Technology develops AI-powered carbon accounting tools. According to Wu, the green transition still depends on cumbersome human work that can be automated.

    Zhongtan Puhui builds AI agents that perform a number of carbon accounting tasks, including carbon footprinting and emissions audits. Customers range from China Minmetals Group and DHL to small and medium-sized exporters in the Yangtze River Delta.