The race to dominate artificial intelligence has produced some of the biggest winners in financial markets over the past two years. Chipmakers have seen their share prices soar. Software firms have rushed to add AI products. Private companies have secured funding at valuations that would have seemed difficult to imagine only a short time ago. For many investors, AI has become the defining investment story of the decade. Yet the strongest rallies often attract equally strong warnings.
That is exactly what happened this week after two respected Chinese hedge fund managers argued that enthusiasm surrounding artificial intelligence has moved far ahead of business reality. Their comments have added fresh fuel to a debate already taking place across financial markets about whether investors are paying too much for future earnings that may take years to arrive.
The warnings arrive at a time when AI companies continue attracting large sums of capital while also facing growing questions over rising computing costs, competition and whether enterprise customers will continue spending at the same pace.
Veteran Investors Question Whether AI Valuations Have Gone Too Far
One of the most closely watched warnings came from Wealspring Asset, whose founder Yang Dong is widely remembered in Chinese financial circles for correctly identifying the end of China’s 2007 stock market rally. According to an investor letter reported by Bloomberg News, the firm described AI shares as a “super bubble” and suggested the point where valuations begin falling sharply “may not be far away.”
The language immediately caught investors’ attention, not only because of Yang’s reputation but also because it echoed concerns that have gradually become more common among fund managers this year.
Wealspring manages more than $1.4 billion in assets and argued that many companies supplying AI computing equipment or related services still depend on continuous spending to support expansion. In its view, many businesses have yet to prove they possess lasting competitive advantages capable of supporting their current market values.
The letter reportedly compared today’s investment behaviour with China’s 2015 equity boom, a period remembered for heavy speculation and rapidly rising share prices before a painful correction. According to the firm, investors appear willing to pay increasingly high prices simply because companies are associated with artificial intelligence.
The fund reportedly warned that some of the hottest AI-related shares listed in China could eventually lose more than 80% of their market value if investor sentiment changes sharply. It did not identify any specific companies.
Shanghai Banxia Investment Management Center delivered an even more direct argument. The hedge fund told investors it believes the trigger for an AI market correction has already appeared. Its reasoning centres on the business economics of AI rather than excitement surrounding new products.
According to Banxia, companies building large language models now face mounting pressure from customers questioning the rising cost of AI usage. The firm argued that enterprise clients may become more selective as computing expenses increase and competitors introduce lower-priced alternatives.
The letter reportedly highlighted Anthropic, one of the world’s largest AI developers, suggesting market expectations for its annualised revenue growth may prove difficult to meet. Since Anthropic remains privately held, many of the financial estimates discussed by investors rely on industry reports rather than public financial statements. That distinction matters because outside investors have limited visibility into the company’s financial position compared with publicly traded businesses.
Rising Costs and Tougher Competition are Changing the AI Conversation
The Chinese hedge funds are not alone in focusing on costs rather than technical progress. Over recent months, several reports have pointed to rising expenses faced by companies using advanced AI models. Many businesses initially experimented with AI under pricing arrangements that made adoption relatively inexpensive. As those systems became part of everyday work, billing increasingly shifted towards usage-based pricing, where every request consumes computing resources and generates costs.
Finance departments that previously viewed AI as another software subscription are now discovering that heavy usage can produce bills that grow much faster than expected. At the same time, competition inside the AI industry continues to intensify.
American companies including OpenAI and Anthropic now compete alongside Google, Meta and a growing list of Chinese developers. New models continue appearing at lower prices while attempting to match the reasoning and coding ability of more established systems. That pricing pressure creates a difficult commercial balance.
Companies need massive investment in computing hardware, data centres and specialised chips while also facing customer expectations that AI services should become cheaper over time.
Investors have largely focused on revenue growth during the current AI rally. Some hedge funds are beginning to ask a different question. They want to know whether those revenues will eventually produce profits that justify current valuations. Recent trading suggests that markets are already becoming more sensitive to those questions.
Technology shares have experienced several sharp pullbacks during the year despite continuing enthusiasm for artificial intelligence. Semiconductor companies, which have been among the strongest performers during the AI rally, have also experienced large swings as investors reassess earnings expectations and future spending by cloud computing companies. None of that necessarily means the wider AI industry faces an immediate downturn.
History shows that new technologies often experience periods where investor enthusiasm runs ahead of commercial reality before business models gradually mature. Similar debates surrounded internet companies during the late 1990s, cloud computing during the previous decade and electric vehicle manufacturers more recently.
The present discussion is less about whether artificial intelligence will remain important and more about how much investors should be willing to pay today for profits that may arrive years into the future. There are also reasons some market participants remain constructive despite recent warnings.
Demand for AI computing continues to rise across software development, healthcare, finance, manufacturing and scientific research. Governments are investing heavily in domestic computing capacity, while many companies still view AI as an area where delaying investment could leave them behind competitors. Those trends continue supporting demand for specialised processors, cloud services and enterprise software. Still, the comments from Chinese hedge funds show that enthusiasm is no longer one-sided.
Instead of debating whether AI matters, investors are increasingly debating valuation, pricing power and profitability. Those discussions are becoming just as important as announcements about new models or faster chips.
Whether the recent warnings prove accurate will depend less on headlines and more on future company results. Quarterly earnings, customer spending patterns and the ability of AI firms to balance revenue growth with rising operating costs are likely to receive closer attention in the months ahead.
