Shanghai clarifies IPO path for cash-hungry AI labs racing against US
The Shanghai Stock Exchange (SSE) has clarified rules for unprofitable artificial intelligence model developers wanting to go public, as China’s large language model (LLM) firms scramble for fresh capital in an intense race with US labs. LLM developers can go public on the Shanghai bourse’s Star Mar
By Xinmei Shen

The Shanghai Stock Exchange (SSE) has clarified rules for unprofitable artificial intelligence model developers wanting to go public, as China’s large language model (LLM) firms scramble for fresh capital in an intense race with US labs.
LLM developers can go public on the Shanghai bourse’s Star Market under a set of listing standards that require them to have an anticipated market cap of at least 4 billion yuan (US$591 million), as well as meeting certain criteria in terms of market potential, according to an exchange statement on Wednesday.
The move aimed to support listings of “high-quality” AI firms that had yet to form “a certain scale of revenue”, according to the statement.
AI players allowed to go public under such rules should have launched and operated at scale at least one LLM product, and need to have established “clear commercialisation arrangements”, according to the statement.
LLMs have “emerged as the focal point of global technological competition”, the SSE said in a separate WeChat post on the same day. Companies should have sustained, high-intensity research and development efforts, computing power investments and specialised talent, and be “in urgent need of support from capital markets”, the WeChat post said.
Also on Wednesday, the SSE amended rules for the Star Market to support listings from companies in other fields, including quantum technology, biomedicine, hydrogen and nuclear fusion energy, brain-computer interfaces, robotics, and sixth-generation mobile communications (6G), to “support high-level self-reliance and self-strengthening in science and technology”.
China’s AI start-ups have been ramping up their fundraising efforts amid fierce competition. Hangzhou-based DeepSeek is closing its first-ever external fundraising, raising around 50 billion yuan at a valuation of about 400 billion yuan, the South China Morning Post reported.
Leading AI model developers Zhipu AI and MiniMax, which both listed in Hong Kong in January, have also kicked off plans to go public in mainland China to tap domestic investor appetite for AI-related shares.
MiniMax announced its decision to issue A shares on the Star Market on May 31, followed by Zhipu’s announcement a day later. Zhipu, traded in Hong Kong as Knowledge Atlas Technology, completed the IPO tutoring process, a key step before filing, on Wednesday. MiniMax’s tutoring began late last month, but there have been no updates since.
Zhipu’s latest flagship model GLM-5.2 ranked fourth on the Artificial Analysis’ intelligence ranking, behind only Anthropic’s Claude Fable 5, Claude Opus 4.8 and OpenAI’s GPT 5.5. The company’s Hong Kong share price has increased more than tenfold since its listing.
Overall, funding for Chinese AI firms jumped nearly threefold year on year in the first quarter, when start-ups secured more than 110 billion yuan, according to data from Zero2IPO Research last month.
