China AI and chip IPOs revive onshore market on Beijing's terms

China’s AI and chip companies are driving the country’s strongest onshore technology IPO run since 2023, giving Beijing a cleaner answer to a financing problem that has become strategic rather than cyclical. Mainland technology listings raised US$3.1 billion in the year to June 18, more than five times the level a year earlier, according to Reuters reporting on LSEG data. For regulators, the rebound matters less as a verdict on broad market confidence than as proof that domestic exchanges can still fund the sectors the state wants to keep close.

The queue shows how selective that answer is. Nearly 50 companies applying for offerings in Shanghai and Shenzhen are seeking 126.1 billion yuan, and many of the largest names sit in semiconductors and artificial intelligence, the same report showed. ChangXin Memory Technologies, the memory-chip maker planning a 29.5 billion yuan Shanghai float, is the clearest sign yet that the reopening is being reserved for capital-intensive champions, not handed to the market wholesale.

Bankers and investors read the same tape more cautiously. Demand exists for fresh China tech paper. The harder question is whether it reflects durable private risk appetite or a policy-shaped allocation that works only while Beijing keeps the gate narrow. A second issue follows: mainland issuance is gaining ground, but it is not yet displacing Hong Kong so much as redrawing the division of labour between the two markets.

A selective reopening

Officials at the China Securities Regulatory Commission and the Shanghai bourse are not behaving as if they are reopening the IPO window for everything. The policy priority is future-industry funding, and large language model companies have been handed easier access to the STAR Market, according to Reuters. The distinction matters because the mainland IPO drought of the past two years was never just a valuation story. It was also a regulatory throttle.

Inside that gate, who gets through says more than how many companies line up outside it. Memory, compute and model infrastructure are getting speed. Consumer-facing tech and the long tail of less strategic issuers are not seeing the same signal. Beijing appears to be using the listing market as an industrial-policy tool, one that can recapitalise domestic champions without reopening the speculative excesses that officials spent years trying to suppress.

ChangXin’s planned deal matters beyond its size. A 29.5 billion yuan flotation would be the largest mainland listing this year, and it would place a chipmaker at the centre of the market’s revival. For policymakers, that is a cleaner political outcome than a rebound led by property proxies or meme-driven retail names. It channels savings into an area where China still wants more self-sufficiency, more scale and, eventually, less dependence on foreign supply.

An exit valve for private capital

For issuers, the appeal is straightforward. After a long freeze in mainland offerings, AI and semiconductor companies suddenly have a venue that offers fresh capital, domestic branding and a state-friendly route to liquidity. Semafor’s reporting on Zhipu AI’s planned share sale suggests that one of China’s best-known model builders sees listing momentum as more than a one-off trade. If Zhipu can float locally after first targeting Hong Kong, the message to peers is that onshore markets are no longer merely a fallback.

Venture funds and private equity groups have been waiting for that exit path. Li He put the point sharply, saying the acceleration in technology offerings has reopened a route that backers had been waiting on for years. In other words, the mainland IPO market is starting to function again as infrastructure for the private-capital cycle, not just as a scoreboard for state enterprises.

Local floats still come with trade-offs. Domestic listings can narrow the investor base and leave pricing more exposed to policy mood than a broad global book would. They also offer something Hong Kong and the US can no longer guarantee for Chinese strategic tech names: a venue that aligns financing with Beijing’s industrial priorities. For issuers working in memory, inference or foundation models, that alignment may now matter as much as the cost of capital.

Mainland and Hong Kong are splitting roles

Set against last year’s slump, the mainland rebound looks sharp. Chinese technology companies raised only US$2.7 billion through domestic listings in 2024, down from US$15.7 billion in 2023, according to the same Reuters data. Hong Kong, by contrast, absorbed much of the pressure-release valve. Chinese tech groups raised US$6.6 billion there in 2025, showing that offshore funding never vanished. It just carried more of the load while the mainland window stayed tight.

A better reading is that the two venues are starting to specialise. Mainland markets are becoming the preferred home for politically favoured issuers that can deepen domestic liquidity and support national technology goals. Hong Kong remains the bridge venue: more international, more flexible, but also less able to offer the signalling value of a mainland blessing. That does not eliminate Hong Kong’s role. It makes the city a complement rather than the only viable route.

The limit of this recovery shows up in CNBC’s analysis of Asia’s struggle to produce mega-IPOs. China has no shortage of engineers, large domestic buyers or ambitious tech firms. What it has lacked is a consistently open market capable of producing blockbuster exits on commercial terms alone. The present rebound improves the funding picture, but through a narrower filter. That makes it more durable politically, though not necessarily more market-like.

Why investors are buying the policy trade

Investor demand sits inside a broader shift in how global portfolios are approaching China. In a separate Reuters report on China attracting buyers despite the global selloff, Christopher Hamilton argued that China’s place in portfolios is changing from a simple emerging-market growth bet into a diversification trade.

“The role of China in portfolios is evolving from a simple emerging-market growth allocation toward a more nuanced source of diversification.”

• Christopher Hamilton, Reuters interview

Hamilton’s point is subtle but important. Investors do not have to believe China has solved its economic problems to buy into a policy-backed technology cycle. They only have to believe that the state is willing to protect a few strategic channels of growth, and that capital will keep being pointed at them.

DeepSeek’s chip project fits the same pattern. Reuters’ report that DeepSeek is developing its own AI chip puts company-level self-sufficiency beside market-level capital allocation. One funds domestic champions. The other tries to reduce dependence on imported infrastructure, foreign exchanges and, over time, foreign capital. Semafor’s analysis of Beijing’s open-source AI push shows the policy ambition runs beyond fundraising. China is trying to shape the terms of AI competition at the same time it reshapes the financing stack underneath it.

A genuine reopening would look different. If mainland listings broaden from AI and chips into sectors with no obvious strategic privilege, the rebound will start to look like a real market restart. If the queue remains concentrated in state-backed priorities, investors should read the boom for what it is: a targeted mobilisation of equity capital.

For now, the narrower interpretation fits the evidence better. China’s IPO market is reviving, but it is reviving on Beijing’s terms. For global investors and rival exchanges, the message is less about a cyclical recovery than about where China wants its next generation of technology champions to raise money, who it wants holding the paper, and how tightly it plans to keep control of the process.