The unstoppable momentum of AI has not only made value-oriented fund managers uneasy, but also prompted heavyweight managers known for balanced portfolios to voice their opinions.
The latest fund second-quarter reports show that some balanced fund managers are beginning to raise "questions" and engage in discussions about the current AI market trend. They have expressed their views and even doubts regarding the sustainability of industry investment and returns, the market's investment logic and valuations, or the underlying commercial logic of the sector.
Zhao Feng's managed RuiYuan Balanced Value Three-Year Holding Mixed Fund explicitly placed the contrasting logic of some AI industry chain companies on the table in its Q2 2026 report.
Similarly, the RuiYuan Research Selected Balanced Three-Year Holding Mixed Fund (Initiative), co-managed by Dong Chunfeng, Qin Wei, and Wu Fei, also dedicated significant space to discussing the current value distribution within the AI industry chain. Zhang Jialu's managed RuiYuan Hong Kong Stock Connect Core Value Fund stated its operational conditions even more directly.
The common thread in these three RuiYuan fund Q2 reports is not questioning "whether AI is important," but rather more specific investment issues: who ultimately captures the value? When will the investment translate into returns? What valuation standard should investors use for trading?
AI Value Largely Concentrated in Hardware Investment End
The most direct questioning of the existing AI value distribution system comes from the RuiYuan Research Selected Balanced Three-Year Holding Mixed Fund (Initiative).
In the Q2 report of this product co-managed by Dong Chunfeng, Qin Wei, and Wu Fei, it states that as AI investment volume rapidly expands, it has become the biggest variable affecting the macroeconomy and markets. However, their portfolio has maintained a relatively cautious attitude towards selecting related targets, primarily due to doubts about the current value distribution system within the AI chain.
"The current state where value is almost entirely concentrated at the hardware investment end is difficult to sustain in the future. Among hardware companies, investment targets that truly meet our criteria for company quality, business model, and competitive moat are relatively scarce, and the valuation systems of some sub-optimal varieties also significantly deviate from our standards."
Dong Chunfeng and others further used overseas memory companies as an example to analyze the potential risks of the logic of "moving up the AI industry chain, finding bottleneck links, and profiting from price elasticity."
They analyzed that chasing price elasticity may, first, underestimate the stimulus effect of price increases on supply. In fact, the influx of new entrants driven by rising prices and high profits has already appeared in some industries previously considered to have barriers.
Secondly, due to inventory behavior, shipment fluctuations upstream are far greater than downstream. Even a modest revision in capital expenditure expectations can lead to significant price declines due to inventory changes.
Therefore, regarding valuation, the conclusion given by Dong Chunfeng and others is: "Since price increases will bring negative feedback from both supply and demand, compared to volume growth, the valuation corresponding to profits from price increases should be heavily discounted."
Compared to chasing short-term price elasticity of upstream products, Dong Chunfeng and others stated a preference for "varieties that rely on their own competitiveness to enter the core AI supply chain and can enjoy the industry's long-term growth." In Q2, the portfolio significantly increased allocation to "consumer electronics leaders with stable main businesses, solid moats, and clear future AI-related business growth."
Their attitude towards AI is not closed off. The report concludes by stating that the portfolio "always maintains an open attitude towards AI" and will continue to actively seek "varieties that can maintain a favorable position in the future AI value distribution for a relatively long period."
Long-term Demand and Pricing Face Potential Pressure
Zhao Feng first clearly affirmed the long-term industrial status of AI.
In the Q2 report of his managed RuiYuan Balanced Value Three-Year Holding Mixed Fund, he mentioned, "We believe AI is the next-generation infrastructure, the core variable for the next phase of the productivity revolution. AI's impact on productivity and social organization is still in its early stages, and future development space remains vast."
But immediately after, he introduced a constraint: "If the current speed of AI value release cannot keep pace with the speed of capital expenditure and hardware inflation, it will negatively impact the demand growth rate and pricing power of the industry chain."
He also placed this issue within the lifecycle of past technological revolutions for observation. From core technological breakthroughs to the realization of societal productivity, it typically takes 50 to 60 years; each revolution experiences a distribution lag where "benefits first go to capital, then to labor." The end of the installation period almost inevitably sees financial bubbles and crashes. The bursting of the bubble does not end the technology; instead, it initiates the "deployment period" where technological dividends diffuse throughout society through asset repricing.
Back to the current market, Zhao Feng observed that in Q2, the market exhibited extremely divergent trends. The investment enthusiasm for tech stocks driven by the AI industry wave continuously drew funds from other sectors, widening the valuation gap between the two.
On one hand, according to market consensus expectations, the 2027 valuation of many tech companies has already reached several tens or even over a hundred times PE. On the other hand, the EV/FCFE of many traditional industry companies and "AI victim companies" has fallen to single digits or low single digits, meaning that as long as free cash flow can remain stable for a few years, investors can recoup their principal.
He thus concluded: "Apart from AI-benefiting industries, other industries still faced fundamental pressures in Q2, with unclear outlooks. However, considering the gap between a payback period of decades versus years, and the risks of profit forecasts spanning decades versus the relative certainty of forecasts for a few years, we believe there is irrationality in market pricing."
This valuation skepticism is not expressed in isolation.
He also mentioned that although he has been continuously following AI technology since the launch of ChatGPT, unlike the belief in Scaling Law held within the industry, there is a certain time lag in tracking industry development. When related stock price increases consistently exceed his research understanding and require continuously raising profit forecasts to support the stock price, investing in these targets becomes more difficult.
He also stated that a significant portion of the companies that benefited from this round of AI rally do not have high competitive moats. If not for short-term supply-demand imbalances, these companies would find it hard to possess such high profitability; once the supply-demand relationship changes, a decline in their bargaining power is foreseeable.
In his view, relying too heavily on demand forecasts to project a company's future profits is not a method he excels at. First, high-growth demand forecasts carry significant uncertainty. Second, when the target company's competitive moat is not high enough and the industry competitive landscape is not favorable, the conversion of revenue into profit also involves great uncertainty, making it difficult to adequately capture investment opportunities from such industry waves.
However, Zhao Feng did not deny that AI will still bring investment opportunities. He stated that he will also increase efforts to track and research the AI industry, as future industry development will still offer many investment opportunities, with both technological changes and application layers still in their early stages.
Reasonable Valuation Range Only After "Phased Adjustments"
Zhang Jialu's managed RuiYuan Hong Kong Stock Connect Core Value Mixed Fund also paid attention to AI.
The report stated that in Q2, artificial intelligence remained the dominant core theme in the market, its strong capital attraction effect "almost drawing liquidity from various traditional industries." As there are few local AI investment targets in Hong Kong stocks, the Hong Kong market did not become a major beneficiary of this industry wave; instead, it suffered significant "capital drainage" pressure. Continuous capital flows to semiconductor hardware leaders in markets like South Korea and Japan, and related targets in the A-share market, further diverted liquidity from the Hong Kong stock market.
In terms of actual operations, this product, on one hand, continued to invest in value-oriented companies that can provide relatively certain cash flows and have become highly attractive after long-term adjustments. On the other hand, it also screened "hardware enterprises that benefit from the global AI computing power industry chain while still being within a reasonable valuation range" for allocation.
Therefore, operationally, it will continue to adhere to a core of value assets, using their stable cash flows and low volatility characteristics to provide solid foundational returns for the portfolio. Simultaneously, it will closely track the phased adjustments in the AI computing power hardware sector and allocate within a reasonable valuation range. It will maintain a balanced strategy of both offense and defense, striving to consistently create excess returns for holders in a complex and changing market.
For the second half of the year, Zhang Jialu predicts, "The AI industry trend is far from over, but internal differentiation will become apparent, shifting from pure concept-driven momentum to strict scrutiny of performance delivery capabilities."
Therefore, her operational principle is to continue adhering to a core of value assets while "closely tracking the phased adjustments in the AI computing power hardware sector and allocating within a reasonable valuation range."
But "phased adjustments" and "reasonable valuation range" together constitute clear allocation conditions. She is not chasing AI computing power hardware at any price but is waiting for the price and valuation to enter a range she deems acceptable.
It is worth noting that by the end of Q2, Shenzhen Sunwin Communication Co., Ltd and Yangtze Optical Fibre and Cable Joint Stock Ltd newly entered the fund's list of top holdings. This indicates that the product is not merely waiting but has already allocated to some hardware enterprises; however, for subsequent new allocations, the quarterly report still emphasizes phased adjustments and reasonable valuations.
Refocusing Investment Questions on Cash Flow, Competitive Moat, and Entry Price
Judging from these three Q2 reports, RuiYuan Fund is not pessimistic about the long-term prospects of AI.
However, the questions raised by the three products fall on three different aspects.
First, whether the current concentration of almost all AI-created value at the hardware investment end is sustainable. Second, whether the speed of AI value release can keep up with capital expenditure expansion and hardware inflation, and whether there is irrational pricing between assets with payback periods of decades versus those with payback periods of years. Third, even if the long-term industry trend holds, AI computing power hardware also needs to undergo phased adjustments and enter a reasonable valuation range before allocation.
The fact that multiple RuiYuan fund managers are "nitpicking" AI is not a denial of the technological revolution, but rather a refocusing of the investment question back to cash flow, competitive moat, and entry price: who can capture the value created by AI in the long term, how soon can this value materialize, and how many years of growth expectations are already priced into the current valuation.
Comments