Шанхай расширяет пятый стандарт STAR Market: убыточные AI-компании выйдут на IPO

In 2026, Shanghai authorities published a package of measures expanding the application of the STAR Market's "fifth set of listing standards" to companies in artificial intelligence, large language models, and other cutting-edge technologies — under this standard, even loss-making companies can go public. The document was prepared by the Shanghai Municipal Party Committee's Financial Office and the city's Development and Reform Commission.

What's changing on STAR Market

The fifth STAR Market listing standard will be extended to AI companies and "low-altitude economy" businesses — according to 36Kr, regulators are preparing special review rules (审核指引) for listing such companies. The key feature of this standard is that it allows companies without profit to go public if they operate in priority technology sectors.

STAR Market is the Shanghai Stock Exchange's science and technology board, launched in 2019 as China's answer to Nasdaq for tech startups. Expanding the fifth standard makes it even friendlier to early-stage companies that don't yet have stable revenue but have strong technology. Shanghai explicitly names its goal — to increase the capital market's "institutional tolerance and adaptability" toward such issuers.

Which industries will get access

The measures list specific new and future industries whose companies will be helped to list on STAR Market. According to the Shanghai authorities' document, the list includes:

• Artificial intelligence and large models (大模型)
• Embodied intelligence — AI-powered robotics
• Controlled nuclear fusion
• Quantum computing
• Brain-computer interfaces (BCI)
• Low-altitude economy — drones and air taxis

Separately, the document singles out artificial intelligence and the low-altitude economy as directions for which the authorities are preparing their own listing review rules.

"Expand the application of STAR

Market's fifth set of listing standards, increase the capital market's institutional tolerance toward loss-making advanced-technology enterprises," — from the document by the Shanghai Municipal Party Committee's Financial Office.

Why this matters for AI companies

AI startups rarely become profitable at an early stage: training large models and hiring researchers burn cash faster than revenue comes in. STAR Market's fifth standard removes the profitability requirement — a company can conduct an IPO based on its technology and market valuation rather than net profit. For large-model developers, this is a direct channel to domestic capital.

Beyond the standard itself, Shanghai authorities promise to make more active use of the IPO pre-review mechanism (试点IPO预先审阅机制) for quality tech companies and to improve the Shanghai Stock Exchange's communication channels with major investment institutions and professional investors. All of this should speed up and simplify the listing process itself.

What it means

China is strengthening institutional support for direct financing of the AI sector: loss-making developers of large models, robotics, and quantum technologies are being given a path to IPO on the domestic exchange — at a time when access to Western capital for Chinese AI companies is limited.

Frequently Asked Questions

What is STAR Market?

STAR Market is the Shanghai Stock Exchange's science and technology board, launched in 2019 for listing technology companies. It allows more flexible listing rules than China's main boards.

What does the fifth set of listing standards provide?

The fifth standard allows companies that are not yet profitable to conduct an IPO, provided they operate in advanced technology industries. Shanghai is expanding its application to AI, large models, quantum computing, and other priority sectors.

Which AI companies does this affect?

The measures affect developers of artificial intelligence, large language models, and embodied intelligence (AI robotics), as well as companies from related advanced industries — quantum computing, nuclear fusion, and brain-computer interfaces.