NextFin

China Steps Up Defence of Stock Market in Bid to Revive AI Boom

Summarized by NextFin AI
  • China's CSRC ordered investment banks to pause IPOs and blocked large shareholder sales, the most direct supply-side intervention to stabilise the battered stock market and revive the AI rally.
  • The PBOC prepared to inject more than 100 billion yuan ($15 billion) of liquidity via a third swap-facility round, while cutting the buyback relending rate to 1% from 1.25% and making it permanent.
  • The STAR50 index lost about 40% in under four months, erasing roughly $480 billion of market capitalisation, while AI stocks now represent 42% of A+H market value and 57% of trading turnover.
  • The Shanghai Composite climbed 1.1% to 3,638 points on policy headlines, but the CSI300 had slumped 8.5% in October and remains about 20% below June highs, suggesting a pause rather than a reversal.

NextFin News - China's securities regulator has ordered investment banks to pause initial public offerings and suspended large share sales by existing shareholders, the most direct supply-side intervention yet in Beijing's campaign to stabilise its battered stock market and revive a deflating artificial-intelligence rally. The move, reported on November 3, came as the People's Bank of China prepared to inject more than 100 billion yuan ($15 billion) of liquidity into brokers, insurers and fund managers through a third round of its swap facility, and after the tech-heavy STAR50 index of AI and semiconductor names lost about 40% of its value in under four months, erasing roughly $480 billion of market capitalisation.

The combination matters. Choking off new listings while flooding the market with cheap buying power marks a shift from the verbal support and piecemeal measures Beijing has leaned on since the AI-driven rally peaked in the first half of the year. It is an admission that the Chinese AI equity trade - which Goldman Sachs estimates now accounts for 42% of the combined market value of A-shares and H-shares and 57% of their trading turnover, up from 24% and 48% respectively in 2024 - has become too large to unwind without dragging the entire market down with it.

The Intervention: Cutting Supply While Pumping Liquidity

The China Securities Regulatory Commission released a statement late Monday, November 2, setting out a coordinated package with the central bank. On Tuesday, regulators told investment banks to halt IPOs and blocked large stake sales by existing holders. The CSRC said it would work with the PBOC to fully deploy two structural monetary tools: the Securities, Funds and Insurance Companies Swap Facility, whose third batch of bidding was scheduled for Tuesday with 31 institutions applying for more than 100 billion yuan in quota, and the stock buyback and holding-increase relending programme, which is being converted from a temporary measure into a permanent mechanism with its interest rate cut to 1% from 1.25%.

The scale is meaningful but not overwhelming. The 105 billion yuan injected through the swap facility in October 2024 and January 2025 bought only a temporary respite. Buyback-related loan applications have reached about 440 billion yuan since the tools were launched in September 2024 - real money, but small against a market where AI-linked equities alone represent a multi-trillion-dollar capitalisation. The CSRC also said it would soon revise buyback rules to make "defensive repurchases" easier, allowing companies to buy their own shares when they fall below their issue price, and would require index heavyweights, technology leaders and firms that raised excessive funds to set up dedicated repurchase vehicles. A first batch of 18 active exchange-traded funds has raised more than 35 billion yuan as of November 2, and margin-financing collateral ratios are being loosened to release additional leverage.

The market's first response was a relief rally rather than a conviction trade. The benchmark Shanghai Composite Index climbed 1.1% on Monday, its biggest one-day percentage gain since August 27, and stood at 3,638 points on November 2. But the CSI300 had just slumped 8.5% in October, its worst month since early 2016, and China's broad blue-chip index remains about 20% below the highs made in June. A one-day bounce on policy headlines is a pause, not a reversal.

The Bubble That Made Intervention Unavoidable

To understand why Beijing is acting now, consider the trajectory of the AI trade. Since the "DeepSeek moment" in early 2025, when a Chinese startup demonstrated frontier-model capabilities at a fraction of Western cost, nearly $5 trillion in market value has been created in Chinese AI-related public equities through share-price gains and new listings. AI stocks went from 24% of China's combined A+H market capitalisation in 2024 to 42%, while their share of daily trading turnover jumped from 48% to 57%. When a single theme carries more than half of all trading activity, a reversal stops being a sector correction and becomes a systemic confidence problem - and confidence is precisely what Beijing has been trying to restore as it urges household savings out of property and into equities.

The reversal has been severe. The STAR50 index, which tracks AI and chip leaders on Shanghai's technology board, fell about 40% in under four months, wiping out roughly $480 billion. The damage is concentrated in the names that defined the rally. Robotics maker Unitree, whose August initial public offering drew demand roughly 8,000 times the shares on offer, has given back about half its value since trading began. Chip designers MetaX and Moore Threads, which surged 700% and 400% respectively on their debuts, have each lost about half their value since July. Optical-communications makers Zhongji Innolight and Eoptolink Technology are down about 60% from their June-July peaks. Memory producer CXMT has shed more than a quarter of its value in the past month alone.

"The stock market later evolved into a speculative crazy bull," said Beijing-based fund manager Dong Baozhen. "And now we see the bubble bursting."

Valuations have compressed but remain stretched by global standards. The STAR50 trades at around 100 times trailing earnings, down from a peak near 250 times, but still more than three times the roughly 30 times multiple of the Nasdaq 100 - even as the US index sits close to its record high. That gap is the clearest evidence that the Chinese rally was priced for perfection on earnings that have not yet arrived.

Why the Tools May Not Be Enough

The mechanism of Beijing's defence is straightforward: reduce the supply of shares through the IPO pause and blocked stake sales, and increase demand through cheap institutional liquidity and forced buybacks. In theory this lifts prices. In practice the transmission channel runs through confidence, and confidence is the one variable regulators cannot mandate.

There is a deeper structural problem. China's stock market has long functioned as a financing vehicle for the state and for strategic industries rather than as a vehicle for shareholder returns. The IPO pause addresses a symptom - dilution pressure - but not the underlying disease. Companies list to raise capital for capex, often in sectors where overinvestment is already the problem. AI capital spending in China is becoming increasingly concentrated in a small cluster of firms, and analysts have flagged echoes of past overinvestment cycles. Flooding the market with liquidity to support valuations while the underlying capex cycle is still working through excess capacity is a familiar pattern in China's policy playbook: it buys time, but time alone does not create earnings.

The buyback push is more promising but structurally limited. Making "defensive repurchases" easier and requiring heavyweights to set up dedicated funds creates a mechanical bid under the market. Yet Chinese corporate buybacks have historically been small relative to market capitalisation, and a buyback funded by a 1% central-bank loan is a carry trade, not a statement of intrinsic value. It supports the price; it does not prove the business is worth more.

The second-order risk is what the intervention does to the market's long-run function. An IPO pause stores up pressure: companies still need to raise capital, and a backlog of listings will eventually hit the market, potentially all at once when the pause lifts. Meanwhile, the real economy offers limited support - the official manufacturing purchasing managers' index fell to 49.3 in October from 50.1 in September, and the non-manufacturing index slipped to 49.7, both below the 50-point line separating expansion from contraction. Liquidity can lift asset prices, but it cannot manufacture the demand that turns capex into profit.

The Counter-Thesis: Beijing Has More Firepower Than Markets Credit

The strongest case against pessimism is that markets consistently underestimate the Chinese state's willingness and ability to backstop its equity market. The "national team" of state asset managers and sovereign vehicles has intervened before, and the PBOC's swap facility is a genuinely innovative tool that gives institutions liquidity for stock purchases without expanding the central bank's balance sheet through conventional easing. Hao Zhou, a Hong Kong-based analyst at Guotai Haitong Securities, said the latest swap move "sends another strong signal of policymakers' commitment to safeguarding capital market stability and supporting investor confidence."

There is also a valuation-reset argument. With the STAR50 down 40% and the CSI300 down 8.5% in a single month, much of the air has already left the balloon. If the US AI rally remains intact and export curbs do not tighten further, Chinese AI names could re-rate from a lower, more defensible base. The state's shift from emergency response to what the CSRC described as a "long-term institutional framework" suggests this is not a one-off panic move but a sustained support regime. Hong Hao, managing partner and chief investment officer at Lotus Asset Management in Hong Kong, noted the practical effect: "What's happening is now going to put a stop for many of the companies trying to get themselves listed."

The counter-thesis, however, confuses the ability to support prices with the ability to create a self-sustaining bull market. Beijing can prevent a disorderly crash; it cannot manufacture the earnings growth and productivity gains that make equities rise on fundamentals. The falsifying signal for the sceptical view is concrete: if the Shanghai Composite can hold above 3,700 and the STAR50 can stabilise above its October lows for three consecutive months while AI-related earnings revisions turn positive, then the intervention is working as a durable floor. If instead the indices break lower on heavy volume despite the IPO pause and the liquidity injections, the market is telling policymakers that supply management is not enough.

What Comes Next: Three Horizons

Short term (weeks): The relief rally has room to run on policy headlines. The IPO pause removes an immediate overhang, and the swap-facility liquidity gives institutions dry powder. Expect volatility to remain elevated and the bounce to be headline-driven rather than earnings-driven. The most vulnerable names are the AI exporters in Goldman Sachs's basket, which gained 126% from early 2025 through June before falling 23% from their peak during the global AI selloff - they remain most exposed to US restrictions and to a slowdown in Western data-centre spending.

Medium term (three to twelve months): The outcome depends on whether the policy mix converts into earnings. The buyback relending facility, now permanent at 1%, should create a mechanical floor under index heavyweights. But the key variable is AI demand - both domestic adoption and export exposure. Domestic AI companies have held up better than exporters since the correction, according to Goldman's analysis, because they are insulated from US curbs and benefit from policy support. If US restrictions tighten further, the exporter leg remains the weak point.

Long term (structural): This is the decisive question. China's AI equity boom is a cyclical bubble superimposed on a structural shift. The structural part - China's need to build a self-sufficient AI stack under US technological containment - is real and will not reverse. The cyclical part - the 40% melt-up and melt-down in the STAR50 - is a liquidity event that has now largely reverted. The policy response treats the symptom (prices) more aggressively than the cause (earnings quality and capital allocation). Until the market rewards returns on capital rather than strategic narrative, rallies will remain policy-dependent and therefore fragile.

The base case is a stabilised but range-bound market: Beijing prevents a disorderly collapse, but without an earnings recovery the indices lack the fuel for a durable uptrend. The upside case requires AI capital expenditure to translate into profit growth and US technology tensions to ease. The downside case is that the IPO pause merely defers pressure while the macro backdrop deteriorates - with manufacturing and services both contracting in October, liquidity alone has diminishing returns.

China's regulators have drawn a line in the sand for their stock market. The question is not whether they can hold it for a quarter - with the full weight of the CSRC, the PBOC and the national team, they probably can. It is whether a market propped up by suspended listings and central-bank-funded buybacks can ever again be trusted as a barometer of economic reality. That answer will determine not just the fate of the AI boom, but whether China can build the deep, liquid equity culture its economy needs to shift away from debt and property.

Explore more exclusive insights at nextfin.ai.

Insights

What did China's regulator order?

How much liquidity is PBOC injecting?

Why did the STAR50 index drop 40%?

What defines DeepSeek moment in 2025?

How big is China's AI equity trade now?

What risks does an IPO pause create?

Can liquidity create market earnings?

How does the swap facility work?

What drives the STAR50 valuation gap?

Which AI stocks lost half their value?

What role does the national team play?

How does US curbs affect AI exporters?

What is the short term market outlook?

Why are buybacks seen as limited?

What is the long term structural view?

How does PMI data impact the market?

What is the base case for China stocks?

Can Beijing prevent a disorderly crash?

What is the new buyback relending rate?

How did Unitree stock perform recently?

Search, ask or analyze anything...