A prominent Chinese hedge fund manager has apologized to investors after his firm significantly underperformed the market, putting the blame on its refusal to join this year’s AI-driven stock market rally and warning that the sector has entered bubble territory.
In a widely circulated mid-year letter, Yang Dong, executive director of Ningquan Asset Management, acknowledged that the firm’s products had suffered their largest drawdown since launch after sticking to a value-investing strategy rather than buying into surging AI-related shares.
The letter has resonated across China’s investment community, where many fund managers who missed the AI boom have similarly defended their focus on company fundamentals while expressing concern over speculative trading in artificial intelligence stocks.
Yet while Yang’s assessment highlights the risks of excessive speculation, it also underscores a broader debate over whether investors are underestimating the long-term economic impact of AI.
Value strategy misses AI surge
China’s equity market has become increasingly polarized this year. While the benchmark Shanghai Composite Index has only risen by a modest 3.56%, the tech-heavy ChiNext Index has surged 32.73%, and the STAR 100 Index has skyrocketed 51.63%.
AI infrastructure, computing power and related technology stocks have delivered spectacular gains, with many doubling or even tripling in value.
Since the lows reached in 2025, more than 3,000 of China’s more than 5,000 A-share listed companies have doubled in price with the overwhelming majority linked to AI themes. Around 50 stocks have risen more than tenfold.
By contrast, many traditional industries have stagnated or declined, particularly during the second quarter as investors shifted capital aggressively into AI.
Ningquan’s portfolio was positioned almost entirely on the opposite side of that trade. The fund held internet giants, utility companies, high-dividend blue chips, beaten-down property developers and leading solar companies.
From a fundamental perspective, Yang argued the investment case remained sound. Large internet companies possess abundant data, computing resources and strong cash flows while continuing to invest in AI yet still trade at valuations well below comparable U.S. technology companies. Power utilities should also benefit from rising electricity demand generated by AI development, while the property and solar sectors have already undergone years of painful adjustment.
However, those arguments failed to translate into investment returns. Yang wrote that he understood the long-term industrial trends but struggled to explain why similar assets had performed so differently in China compared with overseas markets. While U.S. technology and utility stocks continued rising, their Chinese counterparts remained under pressure.
He argued that chasing highly valued AI stocks would have been irresponsible given the risks to investors’ capital and insisted the firm would continue acting as “patient capital” rather than participating in speculative trading.
Underestimating the AI boom
Yang acknowledged that his biggest mistake had been underestimating the scale of the AI infrastructure rally.
He argued that many AI-related manufacturers lack durable competitive advantages, require continuous capital investment to sustain growth and have been pushed to excessive valuations by surging demand. In his view, the market increasingly resembles the indiscriminate buying seen during China’s 2015 bull market.
The comparison echoes previous investment cycles. During the renewable energy boom, soaring prices for polysilicon and lithium carbonate convinced many investors that exceptional profits would continue indefinitely, only for excess capacity and fierce competition to drive many stocks down by more than 70%.
Yang contends today’s AI boom represents a global asset bubble rather than a uniquely Chinese phenomenon.
He cited Warren Buffett’s preferred valuation measure — total stock market capitalization relative to GDP — noting that the ratio has climbed beyond levels seen during the 2000 dotcom bubble in several major markets. The U.S. ratio now exceeds 240%, compared with 180% at the 2000 dot-com peak. Japan and South Korea have also crossed 200%.
Based on that assessment, Yang predicted that many popular AI stocks in China could eventually lose 80% to 90% of their value, while the growing influence of quantitative trading and rapid information flows could make any future correction both sudden and difficult to anticipate. The crash may come sooner than expected, triggered purely by valuation excess, not by weakening demand or supply gluts, he said.
Bubble and opportunity can coexist
While Yang’s caution reflects a disciplined investment philosophy, dismissing AI altogether risks overlooking the scale of the technological transformation now under way. Many AI-related companies undoubtedly trade at valuations that already assume years of future growth. A sharp correction would not be surprising, and many speculative names may never justify current prices.
History, however, suggests that bubbles and genuine technological revolutions often develop simultaneously.
The collapse of the dot-com bubble destroyed countless internet companies, yet it also laid the foundations for the digital economy and produced many of today’s global technology giants such as Amazon and Google. Investors who remained focused on businesses with genuine competitive advantages ultimately benefited despite the painful correction.
The same dynamic will likely play out with AI. This is not a simple industry cycle but a paradigm shift in computing power, manufacturing, office productivity and entertainment. After the bubble bursts, the truly innovative players — leading computing capabilities, proprietary algorithms, valuable data assets and strong technological expertise will continue delivering long-term growth.
Legendary investor Lin Yuan famously said: “Those who fear high valuations are destined to be poor.” His point was that truly great companies always seem expensive in the short term, but over time, growth absorbs valuation concerns.
AI represents something different: a technological leap, not a cyclical recovery. It will inevitably be messy, with bubbles and busts along the way. But viewed over the next one or two decades, today’s enthusiasm may represent only the first phase of a much broader technological revolution rather than its conclusion.
Source:
Gelonghui
