China’s burgeoning artificial intelligence sector, particularly its software and cloud computing segments, experienced a significant upswing in July 2026, with exchange-traded funds (ETFs) focused on these areas outperforming broader market indices. This surge was largely attributed to the successful launch of Moonshot AI’s Kimi 3 large language model (LLM), a development that has injected fresh momentum into the Chinese tech landscape and offered investors compelling new diversification opportunities within the global AI trade.
The KraneShares CSI China Internet ETF (KWEB), a prominent tracker of Chinese internet companies with substantial AI and cloud exposure, demonstrated remarkable resilience and growth. In July 2026, KWEB significantly outpaced both the Roundhill Magnificent Seven ETF (MAGS), which comprises leading U.S. hyperscale technology companies, and the iShares MSCI China ETF (MCHI), a broader representation of the Chinese equity market. This outperformance underscores the targeted strength of China’s AI ecosystem and its growing influence on global technology trends. While analysts at CFRA (a leading provider of independent investment research) maintain a bullish outlook on U.S. hyperscalers, the advent of advanced open-weight models from China, such as Kimi 3, presents investors with a more diversified and potentially high-growth avenue within the AI and cloud computing investment thesis.
Kimi 3: A Catalyst for Cloud Demand and Operational Efficiency
The release of Moonshot AI’s Kimi 3 on July 16, 2026, marked a pivotal moment in the development of large language models. This 2.8 trillion parameter open-weight model has achieved performance benchmarks remarkably close to leading U.S. models, including Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6. While Kimi 3’s capabilities inherently position it as a competitor to proprietary models developed by Chinese AI firms, its impact is far from solely disruptive. Instead, it is poised to stimulate significant external cloud demand and drive internal AI adoption across various sectors within China.
The competitive landscape for LLMs in China is robust. Kimi 3 directly challenges Tencent’s own Hunyuan model. However, CFRA maintains a "Strong Buy" rating on Tencent, recognizing that the company stands to benefit significantly from the broader adoption and integration of Kimi 3. Tencent has strategically pivoted its FinTech and Business Services segment to focus on external cloud and AI enterprise services, a move that aligns perfectly with the increased demand for AI infrastructure and solutions. Furthermore, Tencent is actively pursuing AI-led monetization strategies across its vast advertising and gaming businesses, leveraging advanced AI capabilities to enhance user engagement and revenue generation.
Alibaba faces a similar dynamic. Its proprietary Qwen model will undoubtedly feel the competitive pressure from Kimi 3. Nevertheless, Alibaba Cloud is positioned to be a major beneficiary of the AI revolution. The increasing demand for AI training and inference workloads, fueled by models like Kimi 3, will translate into substantial growth for Alibaba’s cloud infrastructure services. This symbiotic relationship highlights how competition can paradoxically drive growth in the underlying infrastructure and services that power advanced AI. Both Tencent and Alibaba hold equity stakes in Moonshot AI, providing them with direct financial upside from the success of Kimi 3, further solidifying their strategic positioning.
The proliferation of low-cost, open-weight models like Kimi 3 also presents significant opportunities for Chinese companies with large, tech-enabled consumer franchises. Meituan, a top-five holding within the KWEB ETF, is a prime example. The company can leverage the cost efficiencies of open-weight models to optimize its core food delivery operations. This includes enhancing route optimization for its vast network of riders, improving rider dispatch efficiency, and implementing dynamic pricing strategies that respond to real-time demand. Beyond operational improvements, Meituan can deploy these models to build sophisticated agentic workflows for its consumer-facing applications, leading to more personalized and responsive user experiences. Meituan’s strategic investment in Moonshot AI, through its investment division, further underscores its commitment to harnessing the power of these advanced AI technologies.
NetEase, a leading provider of online PC and mobile gaming, is another beneficiary. CFRA anticipates that NetEase will gain greater clarity and predictive power regarding its upcoming game releases, such as "Sea of Remnants," "Blood Message," and "Ananta." Open-weight models are expected to serve as invaluable tools in the game development and growth process, offering advanced coding assistants and AI-assisted content creation capabilities that can accelerate production cycles and enhance game quality. CFRA also holds a "Buy" recommendation on JD.com, a major e-commerce and logistics services provider and a top ten holding in KWEB. The integration of AI into JD.com’s operations is expected to unlock significant revenue potential by elevating both user and merchant experiences. This includes enhancing its search and recommendation systems, refining its merchant tools for order management and data analysis, and boosting productivity in areas like video content review and software development.
China’s July Gains: A Deep Dive into AI and Software-Enabled Businesses
The impressive price gains observed in Chinese ETFs during July 2026 were not uniformly distributed across the market. Instead, they were predominantly concentrated in mega-cap AI and software-focused funds. KWEB, with its strong emphasis on AI-enabled stocks, surged by an impressive 16.4% in U.S. dollar terms. Similarly, the iShares China Large-Cap ETF (FXI), which tracks large-cap Chinese equities listed on the Hong Kong Stock Exchange and includes a significant weighting in tech-enabled communication services and consumer firms, posted a gain of 15.5%. These performance metrics highlight the market’s recognition of the value being created within China’s advanced technology sectors.
In contrast, the iShares MSCI China ETF (MCHI), which includes mainland China-listed A-shares and B-shares, did not experience the same level of AI-driven uplift in July. This divergence suggests that the market’s enthusiasm was more focused on companies with direct exposure to cutting-edge AI development and deployment rather than the broader, more diversified equity market. The VanEck China Semiconductor ETF (SMHC), which focuses on hardware, experienced a significant downturn, falling by 35% in July 2026. This decline was attributed to a global pullback in semiconductor stocks, driven by concerns over the substantial capital expenditures required by mega-cap firms for AI development and persistent supply chain pressures exacerbated by international trade restrictions.
The rapid ascent of Chinese AI models in the global developer community is a compelling narrative. Data from Open Router, a platform that processes over 20 trillion tokens weekly across hundreds of AI models, provides a clear indicator of developer preferences when provided with access to all major models via a single API. As of June 2026, Chinese models commanded an approximate 60% market share of total token usage on the Open Router platform. This represents a dramatic increase from a mere 1.2% in October 2024, illustrating the accelerated adoption rate. The market share of Chinese models on Open Router saw a substantial leap to 10% in March 2025, following the launch of DeepSeek V3, and continued its upward trajectory, reaching around 50% by April 2026.
The high performance and accessibility of models like Kimi 3 are key drivers of this trend, further accelerating the adoption of Chinese open-weight models and directly benefiting the holdings within the KWEB ETF. The strategic decision by Chinese AI providers to allow downloadable model weights has proven to be a highly effective distribution strategy, fostering widespread experimentation and integration. Furthermore, these models are inherently more cost-efficient to serve compared to their U.S. counterparts. This cost advantage stems in part from the necessity for Chinese providers to develop highly efficient architectures, partly as a response to export controls that limited access to advanced U.S. semiconductor technology. This combination of open accessibility and lower operational costs has significantly eroded the market share of U.S. model providers, which has fallen from over 70% in 2024 to an estimated 30%-40% by mid-2026. It is important to note, however, that U.S. providers continue to capture a substantial portion of the overall revenue generated within the AI model market.
Navigating the Future: Regulatory and Geopolitical Considerations
As China’s AI ecosystem continues its rapid expansion, regulatory oversight by the Chinese government will be a critical variable to monitor. The potential for government intervention, particularly concerning the international accessibility of open-weight models, could introduce new dynamics. Beijing might implement restrictions on downloading these models, which could negatively impact the growth trajectory of companies held within ETFs like KWEB. Historically, state intervention in China has often prioritized national security and data privacy concerns over short-term corporate profitability. Major technology players such as Tencent, Alibaba, and Baidu already operate within a stringent regulatory framework that includes robust security and content monitoring protocols.
Geopolitical tensions between the United States and China remain a significant factor influencing the technology sector. Developments concerning the export of high-performance chips from the U.S. to China, and conversely, the export of critical minerals from China that are vital to the global technology industry, will continue to shape investment landscapes and market strategies. These complex interdependencies create both risks and opportunities, demanding careful analysis and strategic foresight from investors and industry participants alike. The ability of Chinese AI companies to innovate and scale, coupled with their growing market share in open-weight models, suggests a continued evolution of the global AI landscape, with China playing an increasingly central role. The strategic decisions made by both governments and corporations in the coming months and years will be crucial in determining the long-term trajectory of this dynamic and transformative industry.
About the Author
Aniket Ullal, SVP, ETF Research & Analytics, CFRA Research, is a recognized authority in the exchange-traded fund industry. He founded First Bridge Data, a leading provider of global ETF data and analytics that was acquired by CFRA in August 2019. Prior to his entrepreneurial endeavors, Ullal held product management responsibilities for S&P’s U.S. indices, including the widely followed S&P 500 and S&P/Case-Shiller indices, which collectively track over $1 trillion in ETF assets. Ullal is also the author of "ETF Investment Strategies" (McGraw-Hill; 2013) and holds degrees from Northwestern’s Kellogg School of Management and the Indian Institute of Management in Ahmedabad. His expertise provides a valuable lens through which to analyze market trends and investment opportunities within the ETF space.
