As of the end of the second quarter, the total market saw 72 actively managed equity funds surpass the 100 billion yuan mark in assets under management, a significant increase of 40 funds compared to the 32 recorded in the first quarter.

Looking at fund companies, E Fund Management led the pack with 14 products exceeding 100 billion yuan. Yongying Fund and China Europe Fund tied for second place with 7 such funds each, followed by Xingzheng Global Fund with 4. Fullgoal Fund, GF Fund, Huashang Fund, ChinaAMC Fund, and Invesco Great Wall Fund each held 3.

Performance divergence was equally stark. In the first half of the year, market style distinctly favored hard tech and advanced manufacturing. By the end of Q2, Founder Fubon Core Advantage A led the rally with a gain exceeding 183%, with as many as 29 of the 100-billion-yuan funds achieving double-digit returns.

However, following the market pullback since July, only four funds have maintained their double-digit gains as of July 21st. Orient Artificial Intelligence Theme A now leads with a 124% increase, while Founder Fubon Core Advantage A has retreated to a 91.95% gain, highlighting the high volatility characteristic of the technology sector.

On the other hand, funds heavily weighted in consumer, healthcare, and traditional blue-chip sectors have underperformed. E Fund Blue Chip Selected, managed by Zhang Kun, has fallen 17% year-to-date, while Invesco Great Wall Emerging Growth A, managed by Liu Yanchun, is down 10%. These two former "trillion-yuan star funds" have yet to recover their losses, indicating a prolonged recovery path for the major consumer sector.

Market Consensus: AI Transitions from 'Narrative-Driven' to 'Earnings-Driven'

A review of second-quarter reports reveals that while fund managers share no disagreement on the long-term value of AI, significant divergence exists regarding investment timing, specific segments, and risk perception. The broadest consensus is that AI is no longer a concept-driven speculation but has entered a phase of fundamental earnings realization.

Tang Xiaobin of GF Vision Intelligent Selection pointed out that the investment logic for AI in 2026 has undergone a fundamental shift. With domestic large language models and Agent technology driving explosive growth in tokens, industry prosperity continues to climb.

Jin Zicai of Caitong Growth Preferred reviewed the analytical framework since 2023, noting that starting from Q2 2025, with the explosion of applications like AI Coding, the AI industry's business models have gradually become viable, receiving significant positive market feedback.

Ren Jie of Yongying Technology Intelligent Selection similarly believes the AI industry's fundamentals are in a healthy development stage. The combined Annual Recurring Revenue (ARR) of leading companies has surpassed $100 billion, with business models becoming increasingly clear.

Zheng Xi of E Fund Information Industry Mixed provided a more macro-level judgment from an industry evolution perspective: companies like Anthropic and OpenAI are driving the industry into the Agent era. Large language models are evolving from interactive tools into productivity tools capable of executing complex tasks, advancing global technology from the inference stage to the Multi-Agent complex task stage.

Zheng Xi believes top-tier large language model companies are becoming new global commercial traffic gateways. The existing software and internet industries face explosive demand and industrial restructuring. AI large model enterprises will gradually replace original internet companies as the new leaders of the technology industry. The accelerated commercialization of AI Agents is ushering in a boom cycle for ARR, significantly improving return on investment and bringing forward the profitability inflection point for large model companies. This directly strengthens the investment drivers for global data centers and significantly boosts industry chain prosperity.

Zhang Mingxin of Huashang Zhiyuan Return has observed the same trend. He noted that in Q2, the AI commercial closed-loop accelerated. With further improvements in model capabilities, Agents are expanding from single-scenario applications like coding to data analysis, legal, and other industries. The ARR of leading model companies, represented by Anthropic, is growing at a steeper rate, and overseas cloud providers' capital expenditures continue to be revised significantly upward year-over-year.

First Area of Divergence: Offense or Defense? The Game of Investment Timing

However, even under the shared premise of being bullish on AI, fund managers have diverged into "offensive" and "defensive" camps regarding investment timing.

The offensive camp is represented by Jin Zicai of Caitong Fund. He believes AI demand has led to an increasing number of tight supply chain segments, from optical chips and fiberglass cloth to MLCCs and PCB upstream materials. Consequently, he has further concentrated positions in these "tight segments," anticipating the tightness may persist for a considerable time and aiming to capitalize on market fluctuations.

The defensive camp is represented by Ren Jie of Yongying Fund. He views the current supply pace of the computing power industry chain as slightly ahead of downstream planning, creating periodic supply-demand mismatch pressure. Therefore, his operations are more cautious and balanced, actively adding defensive arrangements and favoring varieties closer to downstream customers with reasonable price expectations, seeking a balance between industry trends and portfolio stability.

Second Area of Divergence: Where to Focus? Choices in Specific Sectors

Divergence is equally pronounced in the choice of specific sectors.

Yongying Technology Intelligent Selection and GF Vision Intelligent Selection focus on configuring computing power infrastructure directions like optical communication and PCBs, believing the construction of computing power clusters for inference demand remains core.

Zhang Mingxin of Huashang Zhiyuan Return also heavily invests in the PCB direction, but with a more specific logic—PCBs are driven by new-generation server architectures, leading to a significant leap in unit value. High-end capacity in segments like CCL and MLCC is experiencing structural tightness, ushering in a price inflection point driven by genuine AI demand.

Zheng Xiaoxi of Southern Information Innovation and Fang Jian of Yinhua Integrated Circuit focus on semiconductor self-sufficiency, heavily weighting equipment, materials, and EDA fields. They see this as the most certain opportunity against the backdrop of continuous domestic wafer fab expansion.

Yan Kai of Orient Artificial Intelligence Theme, Liang Shaowen of Orient Alpha Technology Intelligent Selection, and Wu Hao of Founder Fubon Core Advantage unanimously favor memory chips. They believe the explosion in AI inference demand is simultaneously driving global memory demand, with intensifying supply-demand imbalances leading to sustained price increases and significant earnings leverage for related companies.

Unique Investment Philosophies of Fund Managers

Yan Kai of Orient Fund focuses on the "short board" areas of the semiconductor industry chain, particularly equipment, materials, and components. He believes these segments have high technical barriers and deeply benefit from national support policies, making them core to achieving key technology self-sufficiency.

Ren Jie of Yongying Fund, while firmly optimistic about AI's long-term value, emphasizes investment experience more. As fund size grows, operations become more cautious, strictly controlling portfolio volatility by adding defensive arrangements.

Jin Zicai of Caitong Fund's strategy revolves tightly around the concept of "tightness," further allocating limited positions to tight segments like MLCCs and PCB upstream materials, anticipating prolonged tightness to capture excess returns.

Zheng Xiaoxi of Southern Fund places investments within the context of dual resonance from macro and industry trends. She is optimistic about the global semiconductor upturn driven by AI computing power and memory cycles while firmly bullish on the domestic self-sufficiency trend, believing the continuous increase in domestic substitution rates offers huge growth space for leading domestic semiconductor companies.

Tang Xiaobin of GF Fund believes 2026 is a year testing fund managers' "alpha" capabilities. While various sub-sectors show improvement, broad-based gains are unlikely. Therefore, he has focused on configuring optical fibers & cables, optical communication, computing power leasing, and AI materials with price increase logic.

Fang Jian of Yinhua Fund emphasizes a long-term, heavy commitment to semiconductor localization. The portfolio does not drift with the market, focusing on semiconductor equipment companies with certain earnings. He firmly believes semiconductors are a long-term, promising sector, and only by taking a long-term view and approach can one capture its growth dividends.

Wu Hao of Founder Fubon Fund proposed the view that "storage power" is becoming the hub of system performance. He believes that in 2026, the global tech industry is entering a period of resonant uptrend driven by AI infrastructure expansion and deepening self-sufficiency.

He stated that upstream equipment and materials benefit first with process iteration. Midstream original manufacturers and module factories, focusing on structural supply-demand improvements and multi-year long-term agreements for HBM, DDR5, and enterprise SSDs, are building higher profitability and revenue visibility. Downstream inference-side multi-level storage is seeing comprehensive volume expansion. Driven by both price increases and demand expansion, the entire storage industry chain features both short-term earnings elasticity and medium-to-long-term valuation re-rating characteristics.

Outlook: Volatility and Divergence in Q3, Alpha Ability Faces a Major Test

Looking ahead, Zheng Xi expects the market is highly likely to enter a volatile phase in the third quarter and will focus on varieties with more solid fundamentals and reasonable valuations.

Tang Xiaobin summarized his judgment: In 2026, every sub-sector has reasons worth "sticking to," but likely no single sector can replicate the glorious moment of the "optical module" sector in 2025.

He stated that if 2025 tested investors' ability to capture beta, then 2026 tests alpha capability. The real test of active management ability has just begun.