China’s capital markets are reopening for hard technology, and the numbers are striking. According to Reuters reporting published by Memeburn, mainland IPOs raised approximately $3.1 billion through June 18, five times the pace recorded over the same period last year, driven almost entirely by artificial intelligence and semiconductor listings. Nearly 50 companies have filed applications to list on the Shanghai and Shenzhen exchanges, with a combined fundraising target of 126.1 billion yuan ($18.7 billion).
Three anchor deals define the pipeline: ChangXin Memory Technologies (CXMT) is planning a 29.5 billion yuan offering that would be China’s largest listing of the year; Zhipu AI is targeting a 15 billion yuan raise on the STAR Market; and Baidu’s chip unit Kunlunxin is awaiting approval for a roughly $2 billion flotation in Hong Kong.
The scale and composition of the pipeline reflect a deliberate shift in Chinese equity financing toward strategic technology. The $3.1 billion already raised by mid-June suggests that domestic investors are re-engaging with primary issuance after a period of caution, while the nearly 50 IPO applications signal pent-up supply concentrated in AI infrastructure and model companies. The fivefold year-on-year acceleration is not a statistical quirk, it reflects a genuine change in market conditions, regulatory appetite, and investor conviction about the commercial viability of China’s AI sector.
(Related: After a 65% Rally, China’s AI Stock Boom Faces Its First Serious Reality Check)
CXMT’s 29.5 Billion Yuan Offering Anchors the Semiconductor End
The largest deal in the pipeline belongs to ChangXin Memory Technologies, whose planned 29.5 billion yuan ($4.1 billion) listing would be the biggest onshore offering of 2026. CXMT is China’s leading DRAM manufacturer, and its IPO is a direct readout of Beijing’s determination to build a self-sufficient memory supply chain. DRAM production is capital-intensive, a single advanced fab requires tens of billions of dollars in equipment, and the public markets are one of the few mechanisms capable of mobilizing that scale of capital domestically.
CXMT’s listing also carries strategic significance beyond its size. Memory chips are a critical input for AI training and inference workloads, and China’s dependence on Samsung, SK Hynix, and Micron for DRAM supply has been a persistent vulnerability in its AI infrastructure buildout. A successful CXMT IPO would give the company the capital to accelerate its transition to more advanced process nodes, narrowing the gap with South Korean and US producers that has historically constrained China’s ability to run large-scale AI training clusters on purely domestic hardware.
Baidu’s chip unit, Kunlunxin, adds a new dimension to the pipeline. Unlike CXMT’s memory-focused approach, Kunlunxin designs AI inference chips, the Kunlun series, that directly compete with Nvidia’s data center products in Chinese cloud deployments. Its pending $2 billion Hong Kong listing signals that offshore capital remains relevant for Chinese AI chip companies even as mainland venues absorb the bulk of the year’s largest deals. The dual-track approach, STAR Market for model companies and memory producers, Hong Kong for chip designers with international investor bases, reflects a sophisticated capital strategy.
Zhipu AI’s 15 Billion Yuan Bid Tests Model-Company Valuations
Zhipu AI’s planned 15 billion yuan ($2.1 billion) raise on the STAR Market adds a model-company dimension to a pipeline otherwise dominated by hardware. Zhipu, which developed the GLM series of large language models and operates the Z.ai platform, is one of China’s most commercially advanced AI model companies, with enterprise deployments across finance, healthcare, and government sectors. Its decision to pursue a public listing, rather than continuing to raise private capital, reflects growing confidence that AI software monetization in China has reached a scale where equity markets can underwrite larger rounds.
The 15 billion yuan target is significant in context. It would represent one of the largest STAR Market technology offerings of the year and would establish a public valuation benchmark for China’s model-company sector at a moment when private valuations have been rising rapidly. If Zhipu’s listing proceeds at its target valuation, it will provide a reference point for investors assessing the worth of Moonshot AI, Baichuan, and other private model companies that have yet to test public markets.
What the Pipeline Signals About China’s AI Investment Cycle
The fivefold acceleration in mainland IPO proceeds is not happening in isolation. It follows a period of intense private-market activity, large funding rounds for robotics startups, model companies, and chip designers, that has been building since late 2024. The move toward public listings suggests that some of China’s most advanced AI companies have reached the revenue and operational maturity needed to withstand the scrutiny of public market investors, and that their founders and early backers are ready to begin realizing returns.
For international investors, the pipeline also raises questions about access. Most of the anchor deals, CXMT’s 29.5 billion yuan offering, Zhipu AI’s 15 billion yuan raise, are targeting mainland exchanges where foreign participation is limited by quota and regulatory constraints. The Hong Kong track, represented by Kunlunxin’s pending listing, offers a more accessible entry point, but it captures only a fraction of the total capital being deployed. The structural barrier between China’s onshore AI capital markets and international investors remains largely intact, even as the sector’s scale and sophistication grow.