The AI fervor is once again surging along the Shanghai Bund. On July 18th, the Shanghai World Expo Exhibition Center was bustling with activity as the World Artificial Intelligence Conference was in full swing.
This year's conference attracted over 1,100 exhibiting companies, with more than 3,000 products on display and over 300 making their global debut. The scale and quality of the event have reached a new level, making it a crucial vantage point for observing the progress of technology implementation and identifying opportunities in niche sectors.
As the delegate sent by the Invesco Great Wall Technology Corps, fund manager Liu Lisi transformed into a tech "scout." He connected a chain of observation spanning "technology-industry-capital" across various exhibition booths, covering areas from foundational models and cloud services to computing chips and terminal manufacturing. For Invesco Great Wall, this on-site visit serves as both an immersive survey of the current state of the AI industry and a practical verification of the "AI commercialization closed-loop," aiming to find more solid industrial logic to support future investment positioning.
On-the-Ground Tour of the Entire Supply Chain, Witnessing AI Implementation Results
A significant shift at this year's conference is the move from simply "showing off muscles" to competing on commercial implementation. Liu Lisi's tour route precisely followed the path of "model layer - cloud services - computing chips - terminal manufacturing," effectively stringing together the entire AI industry chain.
At the MiniMax booth, the newly unveiled M3 large model for 2026 was the absolute center of attention. On-site materials indicated that this open-source model, featuring million-level context length and native multimodal capabilities, has already received adaptation support from several overseas chip manufacturers, visibly lowering the deployment threshold.
Moving to the Alibaba Cloud exhibition area, the comprehensive "AI suite" made a strong impression, showcasing everything from underlying chips to upper-layer Agent applications. The layout of the full-stack architecture clearly demonstrated the systematic nature of technology implementation.
In the Moore Threads Intelligent Computing Card display area, Liu Lisi checked out the flagship AI training-and-inference integrated intelligent computing card. According to on-site staff, this product has already been adapted for multiple mainstream large models and can cover full-scenario training and inference.
At the Kunlunxin exhibition area, the focus of the on-site discussion was on the large-scale deployment and engineering capabilities of cloud-based general-purpose AI chips and intelligent computing centers.
The final stop was Luxshare Precision. The AI phone precision modules, AR/VR structural components, and robotic parts displayed at the booth transformed abstract algorithmic capabilities into tangible industrial products, visually presenting the final mile of AI implementation.
Confidence in Continued Industry Development as the AI Frontline Advances Vigorously
After the immersive experience at the World AI Conference, Liu Lisi's most immediate impression was that this year's technical roadmap has shifted from a diverse array of approaches to a convergence on main themes, with a core focus on coding capabilities and intelligent Agents. Furthermore, a significant highlight of this conference was the notable narrowing of the gap between domestic and international large models, with the overall gap shrinking from one year to just a few months. Multimodal models like Seedance, Kling, and MiniMax's Conch are firmly positioned in the global first tier, and the industry's soft barriers also favor China's continued catch-up efforts.
Despite recent significant volatility in the AI sector within capital markets, the industry frontline continues to advance vigorously. Meanwhile, breakthroughs are occurring at multiple points across the domestic independent supply chain. Although the domestic layout started later than overseas, high-quality enterprises have emerged in every segment—chips, models, and manufacturing—potentially forming an internal industry cycle. Additionally, China's manufacturing advantages are prominent. Its robust supply capacity, increasingly accumulated technical capabilities, and R&D cooperation are winning favor from global clients. The long-term accumulation in automation, intelligence, and digitalization will further empower the AI supply chain in the future. These factors collectively form the basis for confidence in the continued forward progress of China's domestic AI industry.
From an investment perspective, there is long-term optimism for the domestic AI independent cycle sector, which is expected to have ample growth space from 2026 to 2028. The focus will be on selecting areas of import substitution and supply constraints, and on positioning in targets with valuation pullbacks and clear industrial logic. Specific segments to watch closely include cloud computing with improving volume and pricing dynamics, and domestic inference GPUs where orders and production capacity are key considerations.
For general investors, Liu Lisi offers two practical suggestions: first, regularly use AI products to build industry understanding; second, continuously track two core commercialization metrics: the monthly changes in annualized revenue of major companies and the market feedback on new models. AI is a long-term race where technological uncertainty can easily trigger stock price volatility. Investors need to allocate based on their own risk preferences and can also rely on actively managed funds with deep supply chain research to capture long-term investment opportunities.
Invesco Great Wall Technology Corps Sees Proliferation of Doubling Funds
Liu Lisi, who conducted this frontline research at the World AI Conference, is a new-generation fund manager within Invesco Great Wall's stable growth-oriented faction. His research covers multiple fields including electronics, machinery, and transportation cycles. He adheres to deep industry research, identifying companies with clear industry trends and solid fundamentals from a medium-to-long-term perspective. His investment judgments are always grounded in the real development patterns and plain facts of the industries.
Liu Lisi currently manages the Invesco Great Wall Growth Trend Fund and the Invesco Great Wall SME & Innovation Fund. In terms of performance, as of June 30th, the A-share class of the Invesco Great Wall Growth Trend Fund achieved net value growth rates of 54.69% and 85.47% over the past six months and one year, respectively. The Invesco Great Wall SME & Innovation Fund achieved net value growth rates of 67.12% and 116.40% over the past six months and one year, respectively, significantly outperforming their respective benchmarks. (Data source: Wind. The benchmark returns for the corresponding periods for Invesco Great Wall Growth Trend were 32.16% and 78.02%; for Invesco Great Wall SME & Innovation, they were 14.99% and 41.37%. Liu Lisi began co-managing Invesco Great Wall Growth Trend on September 16, 2025, and began co-managing Invesco Great Wall SME & Innovation on November 15, 2024.)
Liu Lisi is also a member of the Invesco Great Wall Technology Corps. This team brings together 12 fund managers, including Yang Ruiwen and Dong Han, who are deeply entrenched in the tech sector. They have built a distinctive investment research framework featuring "vertical specialization + cross-industry collaboration," comprehensively covering semiconductors, artificial intelligence, and high-end manufacturing. This has formed a complete research chain spanning hardware infrastructure, software application scenarios, and end-consumer demand. Wind data shows that as of June 30, 2026, 34 actively managed equity funds under the Invesco Great Wall Technology Corps achieved returns exceeding 30% over the past year, with 14 of those products more than doubling their value. (Actively managed equity funds are classified as: ordinary equity funds, partial equity hybrid funds, and some flexible allocation, balanced hybrid funds, as well as ordinary equity and partial equity hybrid funds under the QDII classification.)
The Technology Corps is a microcosm of Invesco Great Wall's investment research strength. Over its 23-year history, the company has consistently built upon its active management capabilities, adhering to the philosophy of "preferring steady streams over sudden surges," and striving to become a leading expert in diversified asset management. Facing the technological wave led by AI, Invesco Great Wall will continue to leverage its platform-based investment research advantages, continuously improve team building, and support investment decisions with solid industry research, aiming to create sustainable returns for investors.
Note: The above-mentioned individual stocks are listed for illustrative purposes only and do not constitute any specific investment advice. The stock market involves risks, and investment requires caution. The views expressed are based on analysis of current market conditions, are time-sensitive, and are for reference only. The past performance of a fund is not indicative of its future results. Fund investment carries risks.
Comments