China’s AI Bubble Has Burst — $480 Billion Wiped Out

C

China spent years building its artificial-intelligence industry into a symbol of technological ambition.

Investors followed.

Companies involved in artificial intelligence, semiconductors, robotics and data-centre equipment became some of the hottest stocks in the Chinese market. Government support strengthened the investment story, while the competition with the United States gave the sector an additional strategic narrative.

Then the numbers started going the other way.

China’s STAR 50 index, which includes major technology and semiconductor companies listed on Shanghai’s STAR Market, has fallen about 40% from its July peak. The decline has erased approximately $480 billion in market value in less than four months.

The collapse has prompted analysts and investors to describe what is happening as the bursting of China's AI bubble.

But the more important story is not simply that Chinese AI stocks have fallen.

It is what the sell-off says about the economics of China's AI strategy.

From AI euphoria to a $480 billion loss

The Chinese AI rally was built on a powerful combination.

Beijing wanted technological self-sufficiency. Investors wanted the next generation of Chinese technology champions. Companies promised exposure to artificial intelligence, robotics and advanced semiconductors.

The result was a huge flow of capital into companies associated with the AI theme.

At its peak, the STAR 50's valuation became extraordinary. The index was trading at nearly 250 times trailing earnings before its decline. Even after the crash, it remains around 100 times earnings, compared with roughly 30 times for the Nasdaq 100.

That means the market has fallen sharply without becoming cheap by conventional standards.

And that is why investors are asking whether the correction is over.

Why Chinese AI stocks are crashing

There is no single event that explains the sell-off.

Instead, several concerns have converged.

The first is valuation.

Investors had pushed AI and semiconductor companies to prices that assumed years of spectacular growth. Once expectations became extremely high, even relatively modest concerns about future demand could produce large declines.

The second is profitability.

China has become extraordinarily good at manufacturing technology products at scale. But that strength can eventually become a problem when many companies enter the same market.

More factories and more competitors can produce more supply than demand.

That creates price competition.

And price competition is particularly dangerous for companies whose valuations already assume rapid profit growth.

One Chinese fund manager quoted by Reuters warned that AI suppliers and chipmakers could eventually face the same price wars that have affected other Chinese manufacturing industries.

The third concern is geopolitics.

Chinese companies increasingly face restrictions on access to advanced technology and foreign markets. The possibility of further restrictions on Chinese equipment used in data centres has added pressure to companies supplying the AI infrastructure industry.

The fourth is simply that investors are beginning to question how much AI demand can actually translate into corporate earnings.

That may be the most important issue of all.

China has built an AI industry. Now it needs customers.

China has invested heavily in artificial intelligence as part of its broader strategy to become a technological power.

But technological capability and commercial profitability are not the same thing.

A country can produce powerful AI models, advanced chips and humanoid robots without necessarily creating companies capable of generating enormous profits.

That distinction is now becoming visible in the stock market.

China's AI-related shares had grown to represent around 42% of the combined market capitalisation of mainland A-shares and Hong Kong-listed H-shares, according to Goldman Sachs data cited by Reuters. They also represented around 57% of trading turnover.

That is an extraordinary concentration.

It means the AI trade had become much more than a niche investment theme.

It had become one of the central drivers of China's stock market.

When AI stocks fell, therefore, the damage spread well beyond a handful of technology companies.

Retail investors are taking the hit

China's stock market is heavily influenced by retail investors, making speculative technology rallies particularly powerful.

The AI boom attracted individual investors looking for the next major technology winner.

Some stocks experienced extraordinary demand.

Unitree, a Chinese humanoid-robotics company, was reportedly 8,000 times oversubscribed by retail investors at its August listing.

Since then, its shares have roughly halved.

Chinese memory-chip company CXMT has also fallen more than a quarter over the past month, while GPU developers MetaX and Moore Threads have lost roughly half their value since July.

The experience illustrates the danger of investing in a theme rather than a business.

Investors may be correct that artificial intelligence will transform the economy.

They can still lose money buying companies at valuations that assume too much future success.

The broader Chinese market is suffering too

The AI sell-off has also exposed weaknesses beyond technology.

China's CSI 300 index, a major benchmark for large Chinese companies, has fallen about 20% from its June peak.

The Guardian, citing Bloomberg, reported that approximately $2.4 trillion had been wiped from China's stock market since May.

That makes the $480 billion AI figure important, but it should not be misunderstood.

$480 billion is the estimated market value erased from the companies represented by the relevant AI and semiconductor index — not $480 billion in cash disappearing from China's economy.

Market capitalisation is the value investors assign to shares. When stock prices fall, that implied value falls as well.

The distinction matters because the headline number can otherwise make the event sound like a direct economic loss of the same amount.

The underlying message, however, remains serious.

Investors are rapidly reducing the price they are willing to pay for China's AI story.

Beijing's AI strategy faces a new test

The stock-market collapse is particularly awkward for Beijing because China has been trying to create a healthier and more stable domestic equity market.

The government has encouraged households to move more of their savings into equities while attempting to develop a long-term "slow bull" market rather than another cycle of spectacular rallies followed by crashes.

The AI boom appeared to fit that strategy.

Technology companies could raise capital.

Investors could participate in China's technological development.

And Beijing could direct private capital toward strategic industries.

But speculative bubbles create the opposite outcome.

When valuations rise too far, the eventual correction can destroy investor confidence.

That could make Chinese households more reluctant to put their savings into equities — precisely the opposite of what Beijing wants.

The AI bubble is not necessarily the AI industry

There is an important distinction between a bubble in AI stocks and a failure of artificial intelligence itself.

China's AI industry is not disappearing.

The country remains one of the world's largest technology markets, has a large engineering workforce and continues to invest heavily in artificial intelligence, semiconductors, robotics and computing infrastructure.

The problem is valuation.

A company can have an excellent technology product and still be an overpriced stock.

That is the lesson investors learned during previous technology booms.

The internet changed the world.

Many internet stocks still collapsed.

The smartphone transformed communications.

Not every smartphone manufacturer became a global giant.

Artificial intelligence can similarly transform the economy while investors discover that some of the companies associated with the transformation were priced far above their realistic future earnings.

China versus the United States

The timing of China's AI correction is particularly interesting because the United States is experiencing a much smaller pullback in AI-related stocks.

The Nasdaq has fallen around 6% from its recent record, compared with the roughly 40% decline in China's STAR 50.

That difference does not mean American AI stocks are immune.

It means the two markets entered the AI boom from different positions.

China's rally became closely linked to government policy, national technological self-reliance and the expectation that domestic companies would become the country's answer to Nvidia, OpenAI and other American technology leaders.

The United States, meanwhile, has a group of companies already generating enormous revenues from cloud computing, advertising, chips and enterprise software.

That gives American investors more established earnings against which to measure AI expectations.

China's problem is that some of its AI companies are being valued primarily for what they might become.

The market is now asking them to prove it.

The Nvidia problem

The semiconductor industry illustrates the challenge.

Companies building AI chips and data-centre components have benefited from the global explosion in demand for computing power.

But advanced AI hardware is an extremely competitive business.

If multiple companies develop similar products, customers can demand lower prices.

If governments restrict access to foreign markets, companies lose potential customers.

And if AI companies slow their infrastructure spending, chip demand can weaken.

That creates a chain reaction:

AI expectations drive semiconductor valuations.

Semiconductor valuations drive investment.

Investment creates more capacity.

More capacity creates competition.

Competition can reduce prices and profits.

The stock market is now beginning to price in that possibility.

China may have built too much, too quickly

This is perhaps the most uncomfortable question for China's AI strategy.

China has repeatedly used industrial policy to build strategic industries at extraordinary speed.

The approach can work.

China became a manufacturing superpower partly by building enormous production capacity and developing domestic supply chains.

But the same strategy can create overcapacity.

Solar panels provide one example. Electric vehicles provide another.

The AI industry could face a similar problem if too many companies pursue the same opportunity simultaneously.

The country could end up with impressive technological capacity but insufficient profits.

For investors, that distinction is crucial.

A factory producing something the world wants is valuable.

A factory producing something everyone can produce is much less valuable.

A warning for the global AI boom

China's experience should matter to investors outside China.

The global AI industry is experiencing an enormous investment boom.

Companies are spending hundreds of billions of dollars on data centres, chips, electricity and AI systems.

Investors are assigning extraordinary valuations to companies expected to benefit from that spending.

That does not necessarily mean a global AI bubble is about to burst.

But China's experience provides a warning.

Technology demand does not automatically equal shareholder returns.

Investors eventually need to know who captures the profits.

Will AI companies capture them?

Will chipmakers capture them?

Will cloud providers capture them?

Will electricity producers capture them?

Or will competition push most of the economic value toward customers?

That is the question markets are only beginning to answer.

The next phase of China's AI race

The collapse of Chinese AI stocks does not mean Beijing is abandoning artificial intelligence.

If anything, the opposite is likely.

The government still views AI and advanced semiconductors as strategic technologies.

But the financial environment has changed.

Companies seeking to raise money through stock listings may now face greater scrutiny. Investors may demand stronger revenue, clearer profitability and more reasonable valuations.

Reuters quoted Hong Hao, managing partner and chief investment officer at Lotus Asset Management, as saying the sell-off could put a stop to many companies seeking listings.

That could produce a healthier industry in the long term.

Weak companies may find it harder to raise capital.

Strong companies may acquire competitors.

Investors may become more selective.

And companies may have to demonstrate that their AI products can generate real economic value rather than simply attract speculative money.

The $480 billion lesson

China's AI bubble provides a lesson that extends well beyond China.

The artificial-intelligence revolution may be real.

The investment opportunity may also be real.

But the two are not the same thing.

The STAR 50's 40% decline demonstrates what happens when expectations move faster than earnings.

Approximately $480 billion in market value has been erased, while China's broader stock market has suffered an even larger decline.

The remarkable part is that the STAR 50 remains valued at around 100 times trailing earnings even after the crash.

That suggests the market may not yet have finished repricing the sector.

For China, the next stage of the AI race will therefore be less about convincing investors that artificial intelligence is the future.

They already believe that.

The harder task will be proving that AI can make money.

And that may ultimately be the most important test facing the entire global AI industry.

The technology revolution is accelerating.

The financial reality is catching up.