The debate over an AI bubble has moved from a quiet worry to the centre of global markets, and in China the warning has already turned into a crash. According to Reuters, the mainland STAR50 index of AI and semiconductor leaders is down 40 percent in less than four months, wiping out some 480 billion dollars in market value, and the AI bubble has already burst in China.
The fall in China is not a small dip. It has hit a market that had run very hot for two years, and it has shaken the plans of millions of small investors who bought into the AI boom. The question now is whether the same pattern is starting in the United States, or whether American tech stocks are simply taking a breather after a long run up.
What happened in China
Reuters reports that China’s broad blue chip index is down about 20 percent from highs made in June. The STAR50 index, which tracks the country’s leading AI and semiconductor companies, has fallen much further. A 40 percent drop in less than four months is the kind of fall that marks a burst bubble rather than a normal correction.
Beijing based fund manager Dong Baozhen told Reuters that the market had turned into what he called a speculative crazy bull, and that the bubble is now bursting. His words match the numbers. The rally that lifted Chinese AI stocks had run to higher heights than markets in the West or in other parts of Asia, so the fall from those heights has been steeper too.
There has been no single clear trigger for the selling, according to Reuters. Traders pointed to a mix of high prices, tax bills and the threat that the United States or other nations might ban Chinese equipment in data centres. Each of these worries on its own might have caused a small pullback. Together they were enough to turn a crowded trade into a rush for the exit.
The pain has been widespread because AI stocks had become such a large part of the Chinese market. According to Goldman Sachs, AI stocks made up 42 percent of China’s market capitalisation and 57 percent of turnover across mainland listed A shares and Hong Kong H shares. With so much money tied to one theme, few investors in a market dominated by small savers have escaped the fall.
The crash has also been a blow to Beijing’s plans. China had hoped to encourage its savers to put money into stocks to help fund the AI boom. The selloff has slowed that effort. It has also stalled a push to break with a long pattern of boom and bust and to build what officials call a slow bull market that delivers steady returns.
The US selloff
In the United States the picture is less dramatic, but the warning signs are growing. The Nasdaq is down 6 percent from a record high set three weeks ago. That is a small fall next to the 40 percent drop in the STAR50, but it matters because the Nasdaq had been led higher by the very same AI stocks that are now under pressure.
The clearest warning sign in the US comes from the chip sector. The Philadelphia Semiconductor Index, known as the SOX, fell for eight straight trading sessions. That is its longest losing streak since the 2008 financial crisis. The index is now down 24 percent from its mid June peak.
Another measure tells the same story. The ratio of the SOX to the S and P 500 has dropped to 1.45, its lowest since April 2026. When chip stocks fall faster than the wider market for this long, it suggests investors are rethinking how much they are willing to pay for the companies at the heart of the AI boom.
Who says bubble
Seeking Alpha analyst Damir Tokic said on 26 October that this is an unfolding AI bubble burst. In his view, the falls in China and the United States are part of the same story, and he believes the next test will be the earnings of the big tech companies known as hyperscalers. Their results will be watched closely for signs of whether their huge AI spending plans can be sustained.
Tokic is not alone in his concern. A Bank of America survey of fund managers in October found that the AI bubble was cited as a market risk by 23 percent of respondents. That put it second on the list of worries. The top risk, named by 31 percent of respondents, was a disorderly rise in government bond yields.
The survey result shows that the bubble fear is now mainstream among professional investors. It is no longer held only by a few bears on the sidelines. Almost one in four fund managers in the survey named it as a risk, which is a sharp change from a year ago when AI was seen almost entirely as a reason to buy.
The other side
Yet the case for the other side is strong, and it rests on real money rather than hope. The biggest tech companies are still raising their AI spending, not cutting it. Alphabet raised its 2026 capital spending range to 195 to 205 billion dollars in July. Meta narrowed its range upwards to 130 to 145 billion dollars. These are the largest investment plans in the history of either company.
Nvidia, the chipmaker at the centre of the AI boom, is also still growing fast. Its revenue was up 106 percent to 96.2 billion dollars last quarter, and it forecast 108 billion dollars for the next quarter. Those are not the numbers of a company whose customers have stopped buying.
There is another figure that cuts both ways. The International Monetary Fund counts 1.4 trillion dollars of revenue promises tied to circular AI deals, in which tech companies buy from each other in a loop. None of those deals has unwound yet, according to the IMF. Supporters of the AI trade say this shows the spending is real. Critics say it shows how much of the boom depends on companies funding each other’s growth.
What to watch next
The honest answer to the bubble question is that no one knows yet. A bubble is only clear in hindsight, and the same market can look like a bubble to one investor and a buying chance to another. What is clear is that the debate has changed. A year ago the main question was how high AI stocks could go. Now the main question is how much of the rise was real demand and how much was excitement.
The next clues will come from earnings. If the hyperscalers show that their AI spending is bringing in real revenue, and if they keep their capital spending plans intact, the selloff may prove to be a pause. If they start to talk about slowing down, or if their customers pull back on orders, the falls in China and the US chip sector may turn out to have been an early warning.
For ordinary investors the lesson from China is simple. When one theme grows to 42 percent of a market, a change of mood can wipe out hundreds of billions of dollars in weeks. The AI story is far from over, but the easy part of the trade, when every AI stock seemed to rise, looks finished for now.
