China's AI Bubble Already Burst. Read It as a Stress Test, Not a Forecast.

If you have spent the past year waiting for the AI bubble to pop, China has given you a preview. The STAR50, an index of the 50 largest companies on Shanghai's tech-heavy STAR Market, is down 40% in less than four months, "wiping out about $480 billion in market value," Reuters reported this week in a story headlined "In China, the AI bubble has already burst."

The damage has spread to the wider market. China's blue-chip CSI 300 closed 20.2% below its June peak on October 22, officially entering a bear market, according to Reuters via Business Recorder. "All bubbles inevitably burst," Shanghai fund manager Wang Zhuo told Reuters.

How Did China's AI Stock Bubble Get So Big?

Fast, and mostly on enthusiasm. China's tech rally "accelerated after DeepSeek emerged in early 2025," CNBC reported, as investors bet that Chinese AI models and homegrown chips could compete without American technology. AI stocks now make up 42% of China's market value and 57% of turnover across onshore and Hong Kong-listed shares, Goldman Sachs estimates, according to Reuters.

The new listings tell the story. Robotics maker Unitree's IPO was "8,000 times oversubscribed by retail investors," and its shares have since halved. GPU makers MetaX and Moore Threads jumped 700% and 400% on their first trading days and have lost half their value since July. At the top, the STAR50 traded at nearly 250 times trailing earnings. It is still at about 100 times, compared with roughly 30 times for the Nasdaq 100.

Borrowed money helped. China's margin financing balance hit a record 2.83 trillion yuan (US$416 billion) in May, above the 2015 peak, Caixin reported, although it was a smaller share of trading than in 2015.

Why Is China's Stock Market So Volatile?

Two reasons, and both make it a poor model for markets elsewhere. The first is who trades. Retail investors drive about 90% of daily trading in China, compared with roughly 20% to 25% on the New York Stock Exchange, according to HSBC data cited by CNBC. Individual savers chase momentum and flee together, so moves that would be corrections in New York become crashes in Shanghai. One investor in Hainan told Reuters he bought Unitree because he believed the government stood behind it.

The second is who steers. In China the state is a market participant. Beijing told officials earlier this month to "avoid bubbles, industrial hollowing-out, blind investment and 'swarm-like' rushes into high-tech sectors," Reuters reported. Regulators have cracked down on leveraged and high-frequency trading, and state-backed investors "sold shares during rallies to restrain gains," according to Reuters. Beijing deflated the bubble on purpose, chasing what officials call a "slow bull" market, and it overshot.

"It's a crisis of confidence right now not a crisis in the economy," Shaun Rein of China Market Research Group told CNBC.

Is the Chinese AI Selloff a Warning for Western Markets?

Partly, and this is where the story gets interesting. Reuters itself calls the slump "a warning sign for volatile global markets," and some of the triggers are global. Traders cited high valuations, tax bills and the threat that the US and other countries will ban Chinese equipment from data centers. Jason Hsu of Rayliant told CNBC the biggest factor driving China's downturn is "the U.S. market sell-off," with the Nasdaq down about 6% from a record three weeks earlier.

Western central banks are watching the same risk. The Bank of England warned in September that "the risk of a sharper correction persists" if earnings expectations fall on doubts about the pace of AI adoption. The IMF said this month that "faltering AI investment and profitability could challenge optimistic expectations and trigger a correction in asset prices linked to the AI industry." The fund counted about US$300 billion in equity and US$650 billion in debt exchanged for US$1.4 trillion in revenue commitments across 14 AI deals since 2023.

Still, the mechanics are different. Western AI valuations rest on companies with large existing revenues, institutional investors who hold through dips, and central banks that do not buy or sell stocks to manage sentiment. A 40% crash driven by retail panic and deliberate state cooling does not translate neatly to a market where neither condition holds. 24/7 Wall St. describes China's crash as "a checklist rather than a forecast," and finds the US showing "softer versions" of the same warning signs. That is the most useful way to read it.

What Does It Mean for Canadian Investors?

Less direct exposure than you might think, and more indirect exposure than you might like. On October 27, the same day China's AI crash led the business pages, Celestica jumped 11.9% on strong results and led the TSX to a six-day high, The Globe and Mail reported. "We're getting more optimism around the AI trade in the US," said Raymond James's Michael Dehal.

A new Bank of Canada staff working paper found that AI-sector shocks add to downside risk in Europe, China and "especially the United Kingdom," with a limited contribution in Canada. Canadians are still exposed through US holdings, though. The Bank's 2026 Financial Stability Report noted that tech now accounts for nearly the same share of the S&P 500 as it did at the peak of the dot-com bubble. Anyone with a US index fund in an RRSP or TFSA is far more concentrated in AI than they may realize.

China's crash proves that AI enthusiasm can outrun AI earnings, and that the gap closes painfully. It does not prove the same thing will happen in New York on the same timeline, because China's market amplifies every swing. Use it as a stress test for your own portfolio. If AI stocks fell 40% tomorrow, how much of what you own would go with them?