China’s AI stocks have lost hundreds of billions of dollars in value, even as the country’s artificial intelligence industry continues to advance. The divergence reveals an important distinction between technological progress and investment returns.
China’s artificial intelligence industry is confronting a financial reality that may eventually face the wider global AI economy. After two years of extraordinary enthusiasm, the country’s technology stocks have suffered a spectacular correction, raising questions about the commercial foundations of its AI boom.
According to Reuters, China’s STAR50 index, which includes leading AI and semiconductor companies, fell 40% in less than four months during 2026, erasing approximately $480 billion in market value. Robotics company Unitree lost half its value following its August IPO, while semiconductor manufacturers and AI infrastructure suppliers experienced similarly dramatic declines.
Even after the correction, the STAR50 was trading at approximately 100 times trailing earnings, compared with around 30 times for the Nasdaq 100. Investors had attached enormous expectations to China’s AI industry, and those expectations were being rapidly reassessed.
Yet the correction arrived at a remarkable moment for Chinese artificial intelligence. The country has developed some of the world’s most competitive AI models, created an enormous domestic market for AI applications and established itself as a leader in industrial automation.
The underlying technological progress continues. What is increasingly uncertain is how much of the resulting economic value AI companies and their investors can capture.
The Economics of Cheap Intelligence
China has pursued a distinctive approach to artificial intelligence, emphasising efficiency, low prices and the widespread availability of open-weight models.
Companies such as DeepSeek, Alibaba and Moonshot have demonstrated that highly capable AI can be developed and deployed at substantially lower costs than many competing Western systems. Chinese models have become increasingly competitive globally, while their low prices and open availability have encouraged adoption among developers and businesses.
This approach has been extraordinarily successful in expanding AI usage. According to the China Internet Network Information Center, more than 700 million Chinese people were using generative AI by mid-2026, representing over half the population. AI applications are becoming embedded in everything from ecommerce and advertising to manufacturing and public services.
But the same competitive forces driving adoption are creating commercial challenges.
When companies release powerful models openly, competitors can deploy them without paying substantial licensing fees. Providers compete aggressively on API prices, while consumer applications struggle to persuade users to pay for services that are increasingly available for free.
Research published by Epoch AI in October 2026 estimated that China’s six leading AI companies collectively generated approximately one-tenth of the AI-related revenue earned by OpenAI and Anthropic combined. Some Chinese developers have achieved impressive margins on individual services, but the wider industry faces intense competition, expensive infrastructure requirements and limited pricing power.
The result is an industry where usage and technical capability are growing considerably faster than revenue.
China’s strategy is making AI more accessible and potentially accelerating its adoption throughout the economy. But the more successful companies become at making intelligence abundant and inexpensive, the harder it becomes to generate the profits needed to finance further development.
This is one of the defining economic tensions in China’s AI industry, and it may become increasingly important as AI models continue to improve and their costs decline.
The Investment Problem
The financial implications extend well beyond individual AI developers.
China is undertaking an enormous expansion of its computing infrastructure, supported by technology companies, government investment funds and state-backed financial institutions.
A September 2026 Rhodium Group analysis projected that AI infrastructure spending among 13 major Chinese companies and operators could roughly double to $139 billion in 2026, rising towards $193 billion in 2027.
Yet much of this investment is running ahead of the cash generated by the industry. Alibaba, Tencent and Baidu recorded combined negative free cash flow of approximately $2.4 billion in the first half of 2026 as infrastructure spending increased.
The structure of Chinese AI financing makes the stock market correction particularly significant. While American technology companies have increasingly turned to bond markets to support their expansion, Chinese AI companies depend more heavily on equity investment, bank lending and government-backed financing.
Falling valuations could therefore translate into reduced investment, delayed IPOs and more difficult fundraising conditions.
China also faces a familiar industrial problem: overcapacity. The rush to establish AI computing centres has encouraged local governments and private investors to finance infrastructure sometimes poorly matched to commercial demand. Some facilities have reported utilisation rates of just 20–30%, even as leading AI developers continue to face shortages of advanced computing hardware.
The challenge is deciding where additional capital can produce useful capacity and where investment has already moved too far ahead of demand.
China’s wider economic slowdown compounds these pressures. Weak consumer spending and persistent price competition are squeezing established businesses that help finance technology investment.
The danger is that companies must continue spending heavily to remain technologically competitive while the financial resources available to support that investment become increasingly constrained.
How the Market Correction Could Reshape Chinese AI
The correction could accelerate consolidation across China’s AI industry.
The country supports a large ecosystem of competing model developers, semiconductor manufacturers, robotics companies and infrastructure providers. Many are pursuing similar capabilities while competing on increasingly narrow margins.
Companies with substantial existing businesses, such as Alibaba, Tencent, ByteDance and Huawei, are better positioned to absorb these pressures. They can also benefit from AI indirectly through cloud services, advertising, ecommerce and their existing technology platforms.
Independent AI laboratories face a more difficult position. Without substantial profits from established businesses, they remain dependent on external investment, commercial partnerships or rapidly growing revenue from their models.
Government policy adds another dimension.
Artificial intelligence is central to China’s industrial strategy, particularly its ambition to reduce dependence on American technology. Beijing has strong reasons to continue supporting domestic semiconductor manufacturing, computing infrastructure and advanced robotics regardless of short-term financial returns.
Rhodium estimates that more than 60% of equity investment in China’s AI chip and server companies since 2023 has come from state-affiliated sources. These strategically important industries may therefore enjoy more resilient financing than commercially unproven AI applications.
This could produce an increasingly differentiated industry, with state support concentrated on critical infrastructure and private investors becoming more selective about the companies they finance.
But there is another possibility that deserves attention.
China’s wider economy may benefit enormously from artificial intelligence even as investors in AI companies suffer.
The country already has considerable advantages in manufacturing, robotics and industrial automation. As Digital in Asia has previously explored, China operates more than 30,000 smart factories, installs over half the world’s industrial robots and has integrated AI into more than 30% of its large manufacturing enterprises.
For these businesses, increasingly capable and inexpensive AI could deliver substantial productivity improvements. Manufacturers, logistics companies, software developers and ecommerce platforms may capture significant economic value from technologies supplied by AI companies operating on comparatively thin margins.
Indeed, this may be one of the intended consequences of China’s broader AI strategy. By encouraging competition, open models and inexpensive deployment, China is helping turn artificial intelligence into widely available industrial infrastructure.
The benefits of that approach could be distributed throughout the economy rather than concentrated among a relatively small number of technology companies.
The Value of AI and the Price of AI Stocks
China’s AI industry still faces considerable challenges. US semiconductor restrictions complicate technological development, intense domestic competition compresses margins, and infrastructure investment continues to grow faster than demonstrated commercial demand.
The stock market correction exposes genuine financial vulnerabilities, particularly among companies whose valuations depend on years of exceptional growth.
Yet China’s underlying advantages remain substantial: an enormous domestic market, advanced manufacturing capabilities, a deep research ecosystem and a government committed to technological development.
The experience highlights an important possibility for the wider AI economy. Artificial intelligence can transform industries and create enormous economic value without delivering equally extraordinary returns to every company developing it.
China has already demonstrated how quickly AI can become more capable, accessible and affordable. Its strategy may ultimately prove highly successful at spreading the benefits of artificial intelligence throughout the economy.
Whether the companies making that possible can generate returns sufficient to justify their investment remains a much more difficult question.
Discover more from Digital in Asia
Subscribe to get the latest posts sent to your email.
