China’s CSI 300 Hits Bear Market as AI Fears Lift Treasuries

China’s CSI 300 Index fell as much as 1.7% on Thursday, bringing its total losses from a high in June to 20%, Bloomberg reported on 22 October 2026. Bloomberg says Chinese stocks were “dragged down by the unwinding of the global artificial intelligence trade and the lack of stimulus to provide a driver ahead of a crucial earnings season.” CNBC, citing LSEG data, reports that the index has declined for 10 consecutive weeks.

On the same day, Bloomberg reported that US Treasuries rose as investors sought shelter “after stocks stumbled on renewed concerns about the artificial intelligence trade,” with the 30-year yield touching 5.38%, the lowest since 23 September. By our count, reporters tied ten market moves to AI worries within about a day, while five voices or moves pointed the other way.

Business Pill · FINANCIAL CONDITIONS

A one-minute Business Engineer explainer of financial conditions: how several prices, from bond yields to credit spreads, pull on borrowing at once. It teaches the general idea only and says nothing about any company or market in this piece.

Follow-up, 22 October 2026: AI Doubts: Tesla, Meta, Nvidia Slip; China Stocks Lose $2.4T, with the company-level premarket moves, Bloomberg’s $2.4 trillion figure and Axios’ token-price data.

What Happened in China

A bear market is the usual market label for a fall of 20% or more from a recent high. Finwire says the CSI 300 crossed that line intraday and that it is the index’s first bear market since the selloff triggered by the COVID-19 pandemic in 2020.

Finwire says Chinese AI-linked stocks “came under heavy selling”, listing Z.AI down 9.4%, MiniMax down 6.9% and chipmaker SMIC down 6.1% on the day, with NetEase down 6.7%, Alibaba down 6.1% and JD.com down 4.6%. Bloomberg names optical firms Eoptolink Technology and Zhongji Innolight among the top losers and says a gauge of major Chinese companies listed in Hong Kong dropped as much as 3.2%.

Finwire says the SSE Star 50, which tracks technology-heavy mainland stocks, has lost nearly 40% over the current half-year, and that the ChiNext Index has fallen 36%. It says investors are reassessing whether heavy spending on artificial intelligence infrastructure will generate sufficient returns. CNBC quotes Shaun Rein of China Market Research Group, who expects further selling pressure but calls the selloff excessive: “It’s a crisis of confidence right now not a crisis in the economy.”

What Was Posted on X

Bloomberg’s post on 22 October says China’s mainland-listed shares were on track to enter a technical bear market as waning hopes for substantial stimulus, lackluster earnings and lingering jitters over the global artificial intelligence trade weighed on sentiment. The piece uses Bloomberg’s account as the China row of the dashboard above.

WatcherGuru, a markets news account, posted on 22 October that China’s stock market had officially entered a bear market. It gives no source or figures; the piece relies on Bloomberg, CNBC and Finwire for the CSI 300 numbers.

The AI-Fear Dashboard: Ten Moves, Five Counterpoints

Each row was reported between 21 and 22 October 2026, and in each one the outlet itself names AI as a factor. The selection and the split into moves and counterpoints are ours; the facts in each row come from the source named in it. The list is not exhaustive.

The Other Side of the Ledger

Not every signal points the same way. Bloomberg says shares of the Korean neocloud Elice rose for a second straight day, climbing above their IPO price, “suggesting investors remain committed to the AI trade despite surging industry spending.” The FT reported that Treasury Secretary Scott Bessent was “largely positive about the technological advances” at a closed-door conference.

Blackstone’s Jas Khaira, speaking on stage at TechCrunch’s conference, said the debt behind most data centres is ultimately backed “by the most creditworthy companies in the world,” because most are built for hyperscalers. He also said lenders are charging more: “We’re also seeing higher credit spreads.” Bloomberg says Deutsche Bank’s Christian Sewing pointed to fiscal deficits in countries like France “as being of more concern.”

The Structural Read

As we read it, the dashboard mixes two kinds of signal: prices of AI companies themselves, such as MiniMax and CoreWeave’s debt, and markets that do not trade AI directly, such as Treasuries, copper and a national stock index.

As we read it, the second kind is what is new in this week’s reports. The Bank of Korea, as relayed by FirstSquawk, says “concerns about a potential AI investment slowdown contributed to heightened stock market volatility”.

As we read it, the counterpoints rest on who holds the risk: Blackstone’s Jas Khaira says the debt behind most data centres is backed “by the most creditworthy companies in the world,” while Bloomberg says Elice’s rise suggests investors “remain committed to the AI trade despite surging industry spending.”

Three Implications

FOR AI COMPANIES RAISING MONEY Khaira says lenders are charging more: “We’re also seeing higher credit spreads.”

FOR INVESTORS IN CHINA CNBC says retail investors drive 90% of daily trading in China, according to HSBC data, and Rein expects further selling pressure.

FOR POLICYMAKERS Bloomberg names the lack of stimulus alongside the AI trade as a driver of China’s fall, and the FT reports Treasury Secretary Scott Bessent does not see a “systemic problem” with the AI boom.

The Business Engineer Lens

This story maps onto the Business Engineer framework The AI Capex Map.

The framework argues: “This is not a continuation of the cloud capex cycle. This is a different cycle, with different physics.” Its count of 2026 compute capex includes “$80B China” within a total of about $1.04 trillion.

As we read it, the 22 October reports show markets beyond the spenders pricing that cycle, from China’s stock index to US government bonds. The framework was published in May 2026 and does not discuss these moves.

What Is Not Established

The ten-and-five split is our tally of reports we read in about a day, not a market index. Bloomberg names a lack of stimulus alongside AI as a driver of China’s fall, and CNBC also points to a multiyear property slowdown and weak consumer demand, so AI is one cause among several. We read the Japanese bond item only as a headline, and the Bank of Korea line only as relayed by FirstSquawk.

Bloomberg and Finwire both describe the CSI 300 crossing 20% during the session; the closing level was not available to us. Nothing here shows that one market moved because of another.

Connecting the Coverage

On 20 October, our piece on the neocloud shakeout reported The Information’s account that executives expect only seven to 10 neoclouds to survive. On 21 October, the Yardeni AI Bubble Meter put bubble conditions at 83 out of 100 while saying the bubble was not about to burst. The same day we covered an advocacy report saying Big Tech’s off-book debt is at least $3.1 trillion, and CNBC’s report that CoreWeave debt yields about 15%.

What’s new today, as we read it, is breadth: in reports from a single day, outlets named AI worries alongside moves in a national stock index, US government bonds and copper, not only in AI companies’ own debt.

The Bottom Line

As we read it, AI is now being named as a reason for moves in markets that do not trade AI directly. In reports from 22 October the label was attached to a Chinese bear market, a haven bid in Treasuries, a copper decline and a Korean central bank statement, while remarks by a Treasury Secretary, a Blackstone executive and the head of Deutsche Bank, reported on 21 and 22 October, played the risk down.