China AI Bubble Burst: Why $480 Billion Was Wiped Out

Updated November 1, 2026. Figures reflect reporting through October 30.

Chinese AI stocks have fallen hard, and many investors now ask whether the China AI bubble burst. Reuters reported on October 26 that the STAR 50 index lost about 40% in less than four months. That drop erased roughly $480 billion in market value. The selling did not stop there.

By the close on October 30, the STAR 50 stood at 1,160.84, about 48.5% below its 52-week high of 2,255.25. This is a later market snapshot than Reuters’ October 26 estimate of a roughly 40% decline. The two figures reflect different reporting dates and should not be added together.

The evidence points to a sharp repricing, not the end of China’s AI industry. Stock prices ran far ahead of profits. Now investors want proof that all the AI spending will pay off.

Has China’s AI Bubble Burst?

China’s AI stocks have crashed, but the industry behind them is still spending heavily. A falling stock market and a failing industry are two different things. The fairest description is a major valuation correction with signs of a bubble unwinding.

The bubble label fits part of the story. Prices rose fast, valuations became extreme, and a sharp reversal followed. A bubble usually means prices climb far beyond what companies can earn. By that test, many Chinese AI stocks looked stretched.

Still, a falling market does not prove that every company was overvalued. Future earnings will show how deep the damage goes.

What Is the STAR 50 Index?

The STAR 50 tracks 50 major companies on the STAR Market, a Shanghai Stock Exchange board for technology and innovation firms. Many chipmakers and other tech names tied to AI sit in the index. It is not an AI-only index, though, so its moves reflect more than AI news. Calling it “China’s AI index” would be inaccurate.

How Much Did China’s AI Stocks Lose?

The table below lists the main figures. Each one comes from a different date or starting point, so read them side by side, not added together.

Measure	Figure	Reported
STAR 50 decline	About 40% in under four months	Oct. 26, Reuters
Market value erased	About $480 billion	Oct. 26, Reuters
CSI 300 vs. June highs	Down about 20%	Oct. 26, Reuters
STAR 50 close	1,160.84; about 48.5% below 52-week high	Oct. 30, Investing.com
CSI 300 in October	Down 8.5%, steepest monthly loss since early 2016	Oct. 31, Reuters

The CSI 300 is a broad index of large Chinese companies from many industries. Its slide shows that the pain reached well beyond tech.

What Does the $480 Billion Loss Actually Mean?

The $480 billion is lost market value, not cash that disappeared. Market value, or market cap, is the share price times the number of shares. When prices fall, the total falls with them. No one wrote a $480 billion check, and no company necessarily lost that much cash.

Here is a simple example. A company with 1 billion shares at $100 each is worth $100 billion on the stock market. If the price drops to $60, its market value falls to $60 billion. That $40 billion drop shows what investors are now willing to pay. It is not automatically a $40 billion cash loss.

Investor losses work a little differently. Someone who sells below the purchase price locks in a real loss. Someone who keeps the shares has a paper loss for now. Each person’s result depends on when they bought.

The date matters too. Reuters measured the $480 billion over the sell-off up to October 26, and the October 30 drop came after that. Mixing the two would double count part of the decline.

Why Did China’s AI Stocks Crash?

No single event caused the crash. Reuters described several pressures hitting at once, and traders pointed to tax bills, high valuations and export limits. Each factor below made the others worse.

Valuations Were Extremely High

At its peak, the STAR 50 traded at about 250 times trailing earnings. After the drop, it was still near 100. The Nasdaq 100 traded around 30 times at the time of Reuters’ report.

The price-to-earnings (P/E) ratio compares a stock’s price with its yearly profit. A P/E of 100 means investors pay $100 for every $1 a company earns in a year. That only makes sense if profits grow very fast. When growth slips, prices can fall quickly. Even after a 40% drop, the index was still priced far above the Nasdaq 100, so valuation risk may not be gone.

Retail Investors Piled In

Goldman Sachs figures cited by Reuters showed that AI-linked stocks made up about 42% of the combined value of mainland A-shares and Hong Kong H-shares. They also drove about 57% of trading volume. When so much trading sits in one theme, a sell-off there can spread to the whole market.

Unitree Robotics shows how heated things got. Its stock sale drew retail demand about 8,000 times the shares offered. Some buyers saw AI and robotics as a government priority and treated them as a safer bet. Policy support, however, does not guarantee profits for shareholders.

Profits Did Not Match the Spending

Building AI is expensive. Companies pay for chips, data centers, power and engineers long before they earn much back. Then they compete on price to win customers.

Open-weight models add to that pressure. These models let developers download and adapt the model’s weights, which gives customers cheaper options. Analysts quoted by CNA pointed to the difficulty of making money from them. State Street also noted that cheaper capable models can weaken the pricing power of model sellers. They are one factor among several, not the single cause.

Potential U.S. Restrictions Added Pressure

Advanced chips matter for training and running large AI models. U.S. export controls on advanced chips can limit Chinese firms’ access to them, which can slow AI work.

A separate issue involves Chinese suppliers of data-center parts. On August 4, Reuters reported that the Trump administration was drafting limits on imports of Chinese data-center devices, including new optical transceivers. That was a draft proposal, not an enacted ban, and it did not cover chip exports to China. Even so, the uncertainty can hurt firms that count on overseas data-center orders.

A Weak Economy and Tax Worries Added Pressure

China’s AI stocks did not fall in a vacuum. CNN reported on October 29 that Chinese stock indexes were more than 20% below their June highs. Weak consumer spending and a long property slump were weighing on the market. Bloomberg data cited by CNN showed Beijing putting about $19 billion into the ETF market to steady it.

Traders also named tax bills as a source of selling pressure, according to Reuters. Reuters did not tie the sell-off to one trigger. If households and businesses spend less, AI products may also take longer to earn their money back.

Which Chinese AI Companies Were Hit Hardest?

Robotics firms, GPU designers and data-center suppliers all took heavy losses. These companies sell very different products, so a single label like “AI stocks” hides real differences. The table shows moves reported by Reuters in late October. They are past movements, not current prices.

Company	What it makes	Reported drop
Unitree Robotics	Humanoid and other robots	About 50% since August debut
MetaX	GPUs	About 50% from July levels
Moore Threads	GPUs	About 50% from July levels
CXMT	Memory chips	More than 25% in the month before Oct. 26
Zhongji Innolight	Optical modules	About 60% below June–July highs
Eoptolink	Optical modules	About 60% below June–July highs

Unitree Robotics

Unitree makes robots, including humanoid models. Its stock sale drew demand about 8,000 times the shares on offer. After the August debut, the shares lost about half their value. In its first earnings report after listing, Unitree reported a 70% year-over-year increase in third-quarter revenue, but net profit attributable to shareholders fell 17.4% as research and development costs more than doubled.

The story shows the gap between excitement and results. It does not show that robotics has failed as an industry. A company’s share price and its technology are separate questions.

MetaX and Moore Threads

MetaX and Moore Threads design GPUs, chips widely used for AI computing. MetaX jumped about 700% on its market debut, and Moore Threads rose about 400%. Both have since lost about half their value from July levels.

Optical and Memory Suppliers

Innolight and Eoptolink make optical modules, which move data between servers inside AI data centers. Both stocks fell about 60% from their earlier highs. Eoptolink alone dropped roughly 50% in October. CXMT, a major memory chip maker, lost more than 25% in the month before Reuters’ October 26 report.

How Are China’s AI Companies Paying for the Boom?

Most of the money comes from share sales and bank loans, according to Rhodium Group research. The September 17 study helps explain why the stock drop matters beyond trading screens. It studied 13 listed Chinese firms, including cloud giants, telecom companies, AI labs and data-center operators. The figures below are Rhodium’s estimates, not final company totals.

Spending Is Rising Fast

Rhodium estimates China’s AI infrastructure spending at about RMB 932 billion in 2026, or roughly $139 billion. That would be a 103% jump from the year before. The group expects spending to pass RMB 1.2 trillion in 2027, or about $193 billion. Even so, Rhodium puts China’s total at only about 15% to 20% of U.S. spending.

Free Cash Flow Has Turned Negative

Free cash flow is the cash a business has left after paying its bills and its investments. Rhodium found that Alibaba, Tencent and Baidu together produced about RMB 170 billion in 2025. In the first half of 2026, their combined free cash flow was negative RMB 16 billion.

Those two periods differ in length, so the numbers are not a like-for-like annual drop. Negative cash flow also does not mean these firms are in trouble. Large companies often spend heavily on purpose. It does raise a fair question about how long they can keep it up.

Funding Depends on Stocks and Bank Loans

Rhodium found that Chinese AI firms lean mostly on equity sales and bank loans. That matters now. When share prices fall, selling new shares gets harder and less attractive. Smaller, cash-burning companies may feel that pressure most.

This does not mean every company faces a funding crisis. It does mean that weak stock prices can make growth harder to pay for.

Could the U.S. AI Bubble Burst Next?

U.S. AI stocks could face a correction, but China’s crash is a warning about prices, not a forecast. American and Chinese markets differ in company size, financing and investor mix. U.S. investors are asking the same basic question, though: will AI spending earn enough to justify its cost?

Valuations Are Far Apart

After its drop, the STAR 50 traded near 100 times trailing earnings. The Nasdaq 100 traded near 30 times. The two indexes hold very different companies, so the comparison is rough. It still shows how much more optimism was built into Chinese tech prices.

U.S. Giants Have Bigger Cash Engines

Experts quoted by CNA noted that Microsoft, Alphabet, Amazon and Meta already run large, profitable businesses. That income helps fund their AI plans. Many Chinese AI firms rely more on outside funding.

The cushion is thinner than it was, however. Rhodium estimated that free cash flow at Microsoft, Amazon, Alphabet, Meta and Oracle fell from about $191 billion in 2025 to $14 billion in the first half of 2026. It also said those companies borrowed more. As with the China figures, the periods are different lengths.

The October 30 Sell-Off Shows Real Worry

On October 30, Reuters reported that U.S. stocks fell on fears about the cost of AI spending. The Nasdaq dropped 2.86%, and the S&P 500 fell 1.82%. Intel lost 28.6%, AMD fell 10.9% and Nvidia slid 4.2%. A semiconductor index dropped 7.2%.

One bad day does not make a bubble burst. Earnings, cash flow and customer demand will tell the real story. China’s experience offers one clear lesson: strong demand for a technology does not justify every stock price.

Is China’s AI Industry Failing?

No. Share prices fell, but Chinese companies are still spending heavily on AI infrastructure. A stock chart measures investor mood, not engineering progress. A company can build useful technology while its shares sink.

The crash does expose real weak spots. Prices were high, funding needs are large, and AI services face price competition. Chip access and overseas sales add more uncertainty. Companies now have to show that customers will pay enough to cover all of it.

What Could Happen Next?

Recovery depends on business results, not on AI excitement. If revenue and cash flow improve, confidence could return. If earnings disappoint and funding tightens, more pressure could follow. These are the signs worth watching:

• Earnings: Are AI-related companies growing sales and profits?
• Free cash flow: Can big tech firms fund AI spending without draining their cash?
• Paying customers: Are businesses buying AI products at scale?
• Chip access and trade rules: Can Chinese firms get or make the chips they need, and will U.S. limits on data-center parts become law?
• Funding: Can companies raise money without leaning on rising share prices?

Frequently Asked Questions

Did China’s AI bubble burst?

China’s AI stocks suffered a major sell-off. The STAR 50 fell about 40% in under four months, according to Reuters. That fits a bubble unwinding, but it does not show that China’s AI industry has failed.

What caused the $480 billion crash?

Reuters pointed to several pressures at once. These included very high valuations, heavy retail speculation, doubts about AI profits, tax worries and U.S. export uncertainty. No single event explains it.

What is the STAR 50 index?

It tracks 50 major companies on the STAR Market, a Shanghai Stock Exchange board for technology and innovation firms. Many chip and AI-related companies sit in it. It is not an AI-only index.

Did investors lose $480 billion in cash?

No. The figure measures lost market value. Real losses depend on when each investor bought and whether they sold.

Which companies were hit hardest?

Optical module makers Zhongji Innolight and Eoptolink fell about 60% from earlier highs. Unitree, MetaX and Moore Threads each lost about half their value. Memory chip maker CXMT lost more than 25% in a month.

Could the U.S. AI bubble burst next?

It could, but China’s crash does not prove it will. U.S. tech leaders have larger, profitable businesses. They also face rising AI costs and falling free cash flow, so investors are watching closely.

Is China’s AI industry still growing?

Spending suggests it is. Rhodium estimates AI infrastructure spending in China will jump 103% in 2026. The harder task is turning that spending into steady profit.

The Bottom Line

The China AI bubble burst is mostly a story about prices racing ahead of profits. About $480 billion in market value vanished, and the selling kept going after Reuters’ report. The technology itself is still advancing, but companies must now prove they can earn a return on it. That test applies to U.S. AI stocks too.

Sources

• Reuters, In China, the AI bubble has already burst (October 26, 2026)
• Reuters, China stocks’ October loss (via Business Recorder)
• Reuters, Wall Street closes sharply lower as AI cash burn fears spark tech selloff (October 30, 2026)
• Reuters, Trump administration drafting ban on Chinese data-center devices (August 4, 2026)
• Channel NewsAsia, Has China’s AI bubble burst?
• CNN, Why isn’t China’s AI and tech boom saving its stock markets? (October 29, 2026)
• Rhodium Group, Examining China’s AI Financing (September 17, 2026)
• State Street Global Advisors, China’s 2026 model wave and the repricing of the AI stack
• Reuters, Unitree’s Q3 profit down 17.4% on year in test for China’s robotics sector (October 28, 2026)
• Investing.com, STAR 50 index data (October 30, 2026 close and 52-week range)
• Shanghai Stock Exchange, Q&A on Launching STAR Market 50 Index

You Might Also Like:
CMA USA Syllabus and Real-World Application in MNC Finance Teams
International Transport of Frozen Sperm to Fertility Clinics Abroad