Mutual Fund Q3 Reports: Innolight Retains Top Spot, CXMT Rises to Fourth
nashnova research
At the end of Q3 2026, Innolight held the top mutual-fund position at ¥95.9 bn in market value, while CXMT — listed barely two months — vaulted to No. 4. AI-linked names now fill 32 of the top 50 holdings; legacy blue chips keep fading.
What do the top four holdings look like?
Innolight (中际旭创) stays No. 1: 1,557 funds, combined holding worth ¥95.94 bn. Eoptolink (新易盛) is No. 2 at ¥87.08 bn; WuXi AppTec (药明康德) No. 3 at ¥44.68 bn.
CXMT (长鑫科技) — China's leading domestic memory-chip maker — listed in late July and within one quarter attracted 959 funds and ¥42.85 bn in holdings, landing at No. 4.
This means → mutual funds went almost "all-in on day one" for domestic memory. Concentration is extreme — whether earnings can back it up is the biggest open question.
Where did the old blue chips go?
CATL slipped to No. 7, Tencent dropped to No. 20, Kweichow Moutai fell to No. 22 — all three former "poster stocks" are now outside the top five.
Only one bank stock, Bank of Ningbo, made the top 50, at No. 41.
This reflects a systematic shift in fund taste: capital has migrated from consumer, financials, and platform tech into the AI supply chain, steadily marginalizing legacy core assets.
How dominant is the AI chain?
AI-linked companies account for 32 of the top 50 holdings, spanning optical modules, semiconductor equipment, PCBs — printed circuit boards that connect chips to systems — and chip design.
Breakdown: optical modules (Innolight, Eoptolink); semicon equipment (NAURA, AMEC, JEPT); PCBs (WUS, Shengyi, Shennan); chips (Cambricon, Hygon, SG Micro).
Pharma holds eight seats, including innovative-drug names like Hengrui and WuXi AppTec, plus CXO firms — contract research and manufacturing organizations serving drug companies.
In plain terms = the color of mutual-fund portfolios has flipped from "consumer + financials" to "AI + pharma," and AI's weight dwarfs pharma's.
What were fund managers doing in Q3?
Average equity weight across comparable funds fell to 78.32%, down 3.29 pp from Q2's 81.61%. Balanced funds dropped even more — 3.76 pp.
A style rotation took hold in Q3: value plays like coal and banks outperformed, while the tech-growth leaders that had led earlier pulled back. Managers repositioned accordingly.
Moves diverged: Harvest's Chen Tao added optical-comm and edge AI while trimming memory chips and semicon equipment; Harvest's Wang Guizhong cut AI-compute exposure and rotated into AI applications and edge devices; E Fund's Yang Zongchang added CXO, coal, and metals for broader diversification.
This means → managers did not flee AI — they restructured *within* it, pivoting from upstream hardware toward mid- and downstream applications and edge.
What is the outlook for Q4?
Wanjia's Huang Hai expects A-shares to stabilize and strengthen through volatility; Yongying's Xu Tuo sees a range-bound consolidation, flagging that a Fed rate-hike restart plus European inflation could weigh on risk appetite.
Xinhua's Cai Chunhong anticipates a shift to "high-growth tech + low-valuation recovery" dual drivers; Penghua's Zhang Hua'en stresses stock-picking as the key alpha source, betting domestic-chip penetration will lift semicon-equipment firms' industry standing.
In plain terms = fund managers broadly agree the market won't crash or surge; the real debate is whether returns come from tech momentum or from cheap names catching up.
市场有风险，内容仅供研究参考，不构成投资建议。