Economic Observer reporter Hong Xiaotang and intern Zhang Pengrui

With the disclosure of the second quarter report of public funds in 2026, a fragmented and even slightly torn panoramic view of the capital market has emerged.

According to data from the Ji'an Jinxin Fund Evaluation Center, as of the end of the second quarter, the total size of public funds exceeded the 40 trillion yuan mark for the first time, and the total size of equity funds increased by about 8% month on month, reaching about 9.5 trillion yuan. The equity positions of both equity and hybrid funds have increased, with equity positions of equity funds rising from 89.31% at the end of the first quarter to 89.71%, and equity positions of hybrid funds rising from 70.89% at the end of the first quarter to 72.23%.

Under the surface of steady progress in the overall equity position of the fund, a chip migration has occurred - funds have become fervent supporters of the AI (artificial intelligence) industry chain represented by electronics and communications, while the consumer, pharmaceutical, new energy, and financial sectors, which were once seen as "ballast stones", have been neglected and withdrawn.

This rare K-type differentiation not only reshapes the heavy holdings of public funds, but also triggers a value collision among fund managers regarding industry trend premiums and mean regression common sense.

Group up to compete for funding

Taking stock of the public fund holding structure in the second quarter of 2026, funds are flowing from traditional blue chip mainboards to more resilient hard technology positions.

According to Wind data, in the second quarter of 2026, the manufacturing industry remains the industry with the largest proportion of public fund holdings, with a significant increase of 8.30% in the market value of its holdings compared to the previous quarter, reaching 64.86%.

In sharp contrast, the proportion of financial industry holdings in public fund holdings decreased by 1.83% month on month to 5.87%; Among them, the electronics industry in the manufacturing sector holds a 16.42% stake, surpassing the power equipment and food and beverage industries for the first time in history, and topping the list of the largest public fund holding industry.

This intense conversion between sectors is vividly reflected in the list changes of heavily held stocks.

In the second quarter, the top holdings of active equity funds in A-shares were almost monopolized by AI computing hardware companies, with Zhongji Xuchuang and Xinyisheng occupying the top two positions, followed closely by hard technology leaders such as Cambricon, Dongshan Precision, Northern Huachuang, and Zhongwei Company. Zhongji Xuchuang ranks first in active equity fund holdings with a total market value of 260.528 billion yuan.

From the increase or decrease in holdings of individual stocks by active equity funds in the second quarter, Cambrian's increased holdings ranked first in terms of market value, and semiconductor and electronic component leaders such as Sanhuan Group, SMIC, and Northern Huachuang became the target of capital competition. On the contrary, in the reduction list, Ningde Times became the individual stock with the largest market capitalization in the second quarter. Zijin Mining, Kweichow Moutai and some traditional communication and mechanical equipment leaders suffered from capital withdrawal to varying degrees.

The strategy research team of China Merchants Securities believes that this extreme K-type differentiation is, on the one hand, a reflection of fundamental K-type differentiation; On the other hand, in an environment where incremental funds are limited, funds actively embrace the AI mainline and form a clear diversion to other sectors. Although this structure has strengthened the market mainline in the short term, it has also led to an increase in trading congestion and increased vulnerability.

The Narrative of Ice and Fire

The fragmentation of the market has brought an unprecedented stress test to public fund managers. In the operational analysis of the second quarter report, an invisible divide divides fund managers into two completely different camps: the "optimists" who embrace the technology mainline and the "pessimists" who adhere to traditional values.

Zhang Mingxin, fund manager of Huashang Fund, stated that from an industry perspective, the AI commercial closed loop accelerated in the second quarter, accompanied by further improvement in model capabilities. Agents (intelligent agents) penetrated from a single scenario of coding (artificial intelligence assisted coding) to multiple industries such as data analysis and law. The top model company ARR (annual recurring revenue) grew at a steeper slope, and overseas cloud factory capital expenditures continued to significantly increase year-on-year.

In the operation of future fund products, Zhang Mingxin expressed optimistic expectations for AI. He stated that overall, in the volatile external environment and the determined trend of the AI industry, operations will continue to revolve around the high position of the AI industry chain's prosperity mainline. With overseas computing power as the main allocation direction, adjust holdings according to the changes in segmented industry prosperity, increase allocation to bottleneck directions with tight supply and demand and high confirmation of price increases in the industry chain, and continue to explore structural alpha.

As a fund manager who "stood in the light" in the first half of this year, Jin Zicai continued to express his optimism about the AI industry chain in the second quarter report. However, his investment portfolio has been adjusted, and the focus of his holdings has shifted from the previous optical module and PCB (printed circuit board) to MLCC (multi-layer ceramic capacitor), PCB upstream and other directions. Zhongji Xuchuang, Shengyi Electronics and others have withdrawn from the top ten heavy holdings list of the fund.

Jin Zicai explained his warehouse adjustment logic in the second quarter report: "On the basis of continuing to be optimistic about the trend of the AI industry, we see that there are more and more tense links in the supply chain. Perhaps a quarter ago, we only saw tight links in the EML and DSP chips, fiberglass cloth, HVLP4 copper foil, and other aspects of optical chips. But at present, in order to adapt to the new changes in the supply chain, we will further allocate our limited portfolio to industries with scarce links, including MLCC in passive components, upstream PCB and other scarce material directions. ”

However, on the other hand, deep value investors who adhere to consumer, pharmaceutical, and traditional midstream manufacturing faced severe net asset value drawdown and cognitive challenges in the second quarter.

For example, Xu Yan, a star fund manager at Dacheng Fund, said, "In the past many years, the second quarter that just passed was the quarter with the highest average underperformance among funds under my management

According to the second quarter report, the Dacheng Ruixiang Hybrid managed by Xu Yan saw a 5.71% decline in A-shares in the second quarter, which is 12.34 percentage points lower than the benchmark of a 6.63% increase in performance. Xu Yan stated in the fund's second quarter report released on July 21st that there is no "light" in his fund holdings. Except for a few sectors, the vast majority of stocks experienced significant declines in the second quarter, with a median decline of over 10%.

The second quarter report shows that Dacheng Ruixiang's mixed heavy holdings include targets such as China Mobile, ZTE, China National Offshore Oil Corporation, and Guangzhou Shenzhen Railway. Xu Yan stated that his starting point in investment transactions is enterprise value and safety margin, but in the context of the development of various industries in the second quarter, these two terms are very outdated. In a new era of technological trends, there seems to be an undeniable positive relationship between "technology" and "progress", and those who have not integrated into this trend exude an unprecedented sense of nostalgia.

Coincidentally, Penghua Industrial Select Hybrid Fund also experienced difficult times in the second quarter. The fund manager, Chen Jinwei, stated in the quarterly report that "we went through a difficult period in the second quarter, which resulted in a 'bull market loss' experience for our holders. We deeply apologize for this

Chen Jinwei has conducted a profound analysis of the current extreme differentiation market. He pointed out that mainstream market voices attribute differentiation to fundamentals, but the degree of differentiation has far exceeded the scope that fundamentals can explain. Compared to the global market, the valuations of US stocks and South Korean technology leaders are far from the extreme of A-shares, and the price to earnings ratio of some electronic sectors in A-shares has exceeded 100 times.

A piece of iron sheet cannot guarantee long-term profits just because you call it 'silicon-based iron', nor will it have to suffer long-term losses just because you call it 'carbon based iron', "Chen Jinwei said in the quarterly report, perhaps expressing the voice of many reverse investors. He believes that the development of industries will not necessarily regress to the mean, but this does not mean that the profitability of enterprises and the valuation of the capital market will not regress to the mean.

It is not uncommon for fund managers to express "distress" in their quarterly reports. Jiao Wei, the fund manager of Yinhua Fund, also expressed similar confusion. Jiao Wei has experienced three cycles in the A-share market and refers to himself as a "veteran who has pursued his growth dreams". He believes that whether it is mobile internet, new energy, or AI in this cycle, they are all repeating the same pattern time and time again. "There is no new story to speak of. If the fundamentals remain unchanged, he can do as he pleases, and the gentle breeze blows over the mountains and hills

The above viewpoint, known as the 'Lao Deng narrative', seems to be a protest against the single value system of the market. In the current pricing logic, anything unrelated to AI is considered lacking in growth potential; Even if some companies' profits and dividends remain stable.

Mean regression?

When the proportion of public fund holdings in the electronics sector surges to the highest percentile in history, and when high-performance fund managers bet on the same track, the hidden flow of risk may have already surged deep.

A fund manager from a large public fund institution in Beijing admitted to reporters that the productivity improvement brought by AI big models is real, and the short-term performance explosion of computing hardware companies is also evident to all. But investment pricing always buys' future beyond expectations' rather than 'current good'. When the valuation of high enterprises has already overdrawn the expected high growth of performance in the next three to five years, any marginal small negative may trigger a severe stampede under the high valuation system.

Returning to mid July, the storage sector experienced a collective decline. In less than four weeks, many AI star stocks that performed well in the first half of the year experienced a situation where their stock prices almost halved.

For example, Yang Siliang of E Fund admitted that its managed products do not have an AI sector, citing that the expected return and sustainability of this round of capital expenditures still face high uncertainty, and the market has given higher expectations.

Xie Zhiyu from Xingzheng Global also admitted in the second quarter report that he is enthusiastic about the AI inflation chain that the market is pursuing in the second quarter,

Xie Zhiyu believes that it is very dangerous to give a high forward price expectation and forward valuation for the current scarce links. Referring to the previous round of new energy expansion cycle, the turning point of stock prices often leads to the turning point of fundamental prices. Referring to the development history of many industries in the past, the reduction of infrastructure costs is necessary to promote the popularization of new technologies and applications and the expansion of industrial scale. Excessive upstream price increases will have a backlash on downstream application promotion.

In terms of allocation, Xie Zhiyu still exposes a large position in the AI pan technology field. In stock selection, he pays more attention to the long-term demand outlook and the stability of the competitive landscape, downplays short-term price expectations, and allocates more to relatively low tech stocks, such as wafer manufacturing and domestic large models. In the field of semiconductor equipment and materials, he pays more attention to high-quality individual stocks with overseas logic. In traditional industries, Xie Zhiyu believes that after experiencing an extreme decline in the second quarter, many leading companies with excellent financial reports have more allocation value, while the increasing global competitiveness of innovative pharmaceutical companies has also been overlooked by the market in the indiscriminate downward trend.

In fact, more fund managers will be cautious in their portfolio operations. Yang Zongchang, the fund manager of E Fund, remains optimistic about the prospects of AI, but he stated that he will pay more attention to the pace of commercialization of large models and the risks of fluctuations in artificial intelligence capital expenditures. After a significant increase in the valuation of his holdings, he gradually reduced his holdings of some stocks at the end of the second quarter and began to include high-quality companies in traditional industries that have increased their attractiveness in the medium to long term.

(Author Hong Xiaotang)