The average drawdown of funds that have doubled in value exceeds 45%, what went wrong?

On October 28, the disclosure of funds' Q3 reports officially concluded, and the top ten heavyweight stocks of public offering funds were subsequently released.

Since the third quarter, the volatility of the technology sectors in A-share and Hong Kong stock markets has intensified with obvious pullbacks, and the funds that doubled their yields and dominated the market in the first half of this year have suffered sharp drawdowns.

In the market downturn where both valuable and low-quality assets are dragged down, fund managers have shown significant divergence in response strategies: some choose to stick to their positions against the trend or even increase their holdings in leading large model and computing power enterprises, enduring a deep drawdown of nearly 60%; others keenly perceive marginal changes in the industry, decisively shift large positions to edge-side AI (Artificial Intelligence), high dividend and traditional defensive sectors, and control the drawdown within 20%.

Behind the divergence lies different understandings of different institutions on the sustainability of the technology industry's prosperity, valuation tolerance and economic recovery momentum.

Reporters from The Economic Observer tried to restore the hidden battle of institutional position adjustment under this drastic market fluctuation by tracking the changes of heavyweight stocks in the Q3 reports of public offering funds.

Roller Coaster

The cycle shift of the capital market is more drastic than expected.

Looking back at the first half of 2026, driven by the explosion of the AI industry and the growth of computing power demand, a number of equity funds heavily positioned in the semiconductor industrial chain stood out. Entering the third quarter, the continuous rise in US bond yields overseas suppressed the valuation of global risk assets, coupled with factors such as the excessive rise of technology stocks in the early stage and crowded institutional allocation, the market has seen a sharp pullback.

According to Wind data, as of October 27, more than 200 funds that doubled their yields with a return rate exceeding 100% in the first half of the year have "all suffered losses", and the product with the highest yield is E Fund Supply Reform, with a year-to-date yield of 97.48%.

For this batch of products whose performance once outperformed the whole market, their average interval return since the third quarter is -46.26%, and some products have given back the excess returns accumulated in the early stage.

Wind data shows that as of October 27, Founder Fubon Core Advantage A, which gained a 183.67% return in the first half of the year and won the half-year championship, has a drawdown of 57.30% since the third quarter; followed by Caitong Multi-Strategy Fuxin, which had a 172.94% return in the first half of the year and also failed to escape the market slump, with a 53.13% drop since the third quarter.

Moreover, as of October 27, Soochow Multi-Strategy A became one of the doubling funds with the largest drawdown since the third quarter with a return rate of -60.44%, while Oriental Alpha Technology Selection A and Rongtong Industry Trend A also suffered heavy losses of -58.90% and -58.17% respectively.

On the other hand, some funds have shown strong resilience against declines. E Fund Supply Reform Flexible has a drawdown of 17.06% since the third quarter, becoming the "anti-decline pioneer" among similar products; Harvest Science and Technology Innovation also locked its drawdown in the same period at 26.08% in the volatile market.

Through the details of heavyweight stocks disclosed in the third quarter, it can be seen that behind the difference in drawdowns are different position adjustment paths in the third quarter.

Mirror of Position Adjustment

From the data of top ten heavyweight stocks disclosed by public offering funds in the third quarter, in the face of the overall pullback of technology stocks, fund managers' responses can be roughly divided into two factions: one is the track adherents, who try to find more explosive investment opportunities in the pullback; the other is the balanced rotation faction, which controls drawdown by reducing position concentration and increasing allocation to defensive assets.

As one of the products with severe drawdown in the third quarter, the position adjustment track of Oriental Alpha Technology Selection A is full of adherence to the semiconductor cycle.

The fund's previously disclosed Q2 report shows that it once heavily held leading enterprises in general storage and semiconductor segments such as GigaDevice, Shannon Core, Longsys and Ingenic. Entering the third quarter, despite the huge selling pressure on technology stocks, the fund still adhered to its positions in industry allocation.

The Q3 report shows that the top five heavyweight stocks of Oriental Alpha Technology Selection A are still concentrated in the information technology and industrial fields: it newly added Changxin Technology as the third largest heavyweight stock, and raised Puyuan Semiconductor to the fourth largest heavyweight stock. Although it reduced its holdings of targets such as GigaDevice, Ingenic and Shannon Core, these stocks still occupy heavyweight positions.

However, this high-position adherence in a single track did not get positive market feedback in the third quarter. According to Wind data, among the first-level TMT industries classified by Shenwan, the electronics and communication sectors fell by 32.65% and 28.72% respectively in the third quarter.

Soochow Multi-Strategy A, which also faced deep net value drawdown, and Founder Fubon Core Advantage A, the champion of the first half of the year, still focused their position adjustment logic on the semiconductor and information technology sectors.

Founder Fubon Core Advantage A sold out storage chip stocks such as Demingli in the third quarter, pushed Changxin Technology to the top of its heavyweight stocks, and retained targets such as Ingenic and Jingzhida. Soochow Multi-Strategy A continued to heavily hold GigaDevice and Puyuan Semiconductor. Under the general downturn, the adherence inevitably suffered the damage of the systematic pullback of the sector.

Different from the above products' adherence against the trend, Harvest Science and Technology Innovation and E Fund Supply Reform chose to adjust positions flexibly.

As a technology-themed fund, the top five heavyweight stocks of Harvest Science and Technology Innovation at the end of the second quarter were all semiconductor equipment and manufacturing manufacturers such as Zhongke Test, GigaDevice, Advanced Micro-Fabrication Equipment and Naura Technology. By the end of the third quarter, the fund sharply reduced its positions in high-value pure computing power and semiconductor equipment, and turned to AI applications and edge-side AI with higher performance certainty and lower valuation. For example, Luxshare Precision and Tencent Holdings jumped to the top two heavyweight stocks in the third quarter, and China Telecom with obvious dividend characteristics was also added to its positions.

A series of position adjustments enabled the fund to control its drawdown to 26.08% in the main decline wave of technology stocks since the third quarter, significantly outperforming its performance benchmark and similar products in the same period.

A more thorough rotation is seen in E Fund Supply Reform Flexible. Against the background of a slight decrease in overall stock position, the fund stepped out of the technology sector and significantly increased its allocation to CXO (Pharmaceutical Research and Development Outsourcing), coal, non-ferrous metals, and mechanical equipment industries in the third quarter. The diversified and balanced allocation portfolio made the fund only have a 17.06% drawdown since the third quarter, performing steadily in the former doubling fund camp.

Looking Ahead

Facing the nearly 60% drawdown since the third quarter, Li Jin, fund manager of Rongtong Industry Trend A, still showed firm confidence in "industry trends with epochal significance" in the Q3 report. Li Jin admitted that the large drawdown in the third quarter was due to the heavy allocation to the upstream links related to price increases. After realizing that the sustainability of price increases is not strong, he has shifted his positions to the downstream links.

Li Jin's logic for being optimistic about the market outlook lies in low valuation and real demand. Li Jin believes that although the development of the artificial intelligence industry will not be smooth sailing, it has huge growth space in the long run. On the one hand, leading A-share computing power enterprises have a valuation of about 10 times PE (Price-to-Earnings Ratio) next year, and leading Hong Kong stock internet enterprises have a valuation of less than 15 times PE, both of which are at relatively low levels from the valuation perspective; on the other hand, the data call volume of application scenarios such as coding and agents (intelligent agents) is growing exponentially, and the quarterly reports of relevant cloud manufacturers show that demand is growing rapidly.

Li Jin said that as the time enters the Q3 report disclosure period, the computing power sector is expected to usher in a performance-driven rising market. In terms of the visible profit ceiling that the main leading computing power enterprises can reach in the foreseeable future, there is still much room for their market value to rise. At the same time, after autonomous driving accumulates a large amount of video training data, it is expected to cross the critical point of industrial development in the next 1-2 years; driven by the low cost and high-efficiency iteration of China's supply chain, the growth speed of the robotics industry has increased significantly, and breakthroughs are also expected in the follow-up.

Wu Hao, fund manager of Founder Fubon Core Advantage, said in the Q3 report that after the adjustment in the third quarter, transaction congestion and leverage levels have dropped significantly, and valuation pressure has been released periodically. He judged that structural opportunities will still be greater than index opportunities, structure is more important than position, and we should focus on selecting directions with strong earnings certainty and verifiable prosperity for layout; at the same time, pay close attention to the marginal changes in external disturbances such as the implementation pace of stable growth policies and overseas liquidity.

Looking forward to the market outlook, Wu Hao believes that the storage price increase trend will continue in the fourth quarter with a slower rising slope. In 2027, storage will account for more than 50% of cloud capital expenditure, and the HBM (High Bandwidth Memory) market size is expected to double. On the supply side, HBM4 will occupy advanced production capacity, and the release rhythm of new supply is slow. The tight supply-demand pattern may continue until the end of 2027, and the industry driving force will shift from the inventory cycle to AI structural supply and demand. The risk lies in the release of new production capacity in the second half of 2027 and the narrowing of the upside space for earnings revision. Therefore, at the allocation level, we should shift from general storage elasticity to HBM and high-end enterprise-level SSD (Solid State Drive) with long-term agreement lock-in and high visibility, and avoid game trading in the consumer end. At the same time, pay attention to the deterministic opportunities on the equipment side brought by the increase of capital expenditure and capacity expansion of domestic storage manufacturers.

Compared with the optimists, the Q3 report of Harvest Science and Technology Innovation shows a more prudent attitude. Wang Guizhong, the fund manager, pointed out that marginal changes have been observed in the third quarter: the ARR (Annual Recurring Revenue) growth rate of overseas leading model companies has slowed down marginally, and the penetration of AI from developers to white-collar workers still takes time.

At the same time, Wang Guizhong expressed his concern about valuation: "In the early stage, some computing power stocks rose sharply, and the future 1-2 years' earnings forecasts implied by the market consensus expectations are already at a relatively high level, so the valuation tolerance space for performance fulfillment has narrowed accordingly."

Based on this judgment, Wang Guizhong reduced part of his positions in AI computing power, and increased the allocation of AI application and edge-side AI targets.

In the medium and long term, Wang Guizhong said that he is still optimistic about the development of the AI industry. The penetration and diffusion of any new technology needs to go through the process of continuous cost reduction and continuous popularization of applications, and AI is in this stage; the whole society is accelerating the application of AI to improve production efficiency, which is the general trend. But when it comes to stock investment, we still need to comprehensively evaluate factors such as the outbreak rhythm of new AI scenarios in the next 1-2 years, the performance forecast and valuation level of the computing power sector, and the direction of structural new technological breakthroughs, and dynamically adjust the AI-related position allocation and individual stock selection.

Yang Zongchang, fund manager of E Fund Supply Reform Flexible, said about the position adjustment in the third quarter that he continued the position adjustment strategy at the end of the second quarter at the beginning of the third quarter, continued to adjust the portfolio, slightly reduced the overall stock position, and greatly optimized the held stocks.

At the same time, in the Q3 report, he emphasized more on the adaptability to the overall macro environment. Yang Zongchang pointed out that while maintaining attention to the large model technology and commercialization progress, he increased research on high-quality companies in traditional industries, and short-term headwinds may bring good prices and buying opportunities instead.

This article is from the WeChat Official Account "The Economic Observer", Author: Hong Xiaotang, published with authorization from 36Kr.