Qiming Venture Partners, one of Asia's most active artificial intelligence investors, has deployed 12 billion yuan (approximately $1.8 billion USD) across more than 100 AI projects, positioning itself at the center of China's AI boom. The firm's portfolio spans foundation models, embodied intelligence, AI infrastructure, and applications, with notable investments in Kling AI, a video generation platform, alongside robotics companies and chip makers. This strategic positioning offers a rare window into which AI technologies venture capitalists believe will define the next decade.

On July 19, 2026, at the World Artificial Intelligence Conference in Shanghai, Qiming Venture Partners released its Top 10 AI Outlook for 2026, marking the fourth consecutive year the firm has published this industry forecast. The announcement underscores how the firm has evolved from early-stage investor to a bellwether for AI trends across Asia. Managing Partner Zhou Zhifeng reflected on the firm's journey, noting that the current AI era demands both historical perspective and philosophical thinking amid the noise of rapid innovation.

What Makes Embodied Intelligence the Biggest Opportunity?

Zhou Zhifeng identified embodied intelligence, the field of physical robots powered by AI, as the area that has "killed the most of my brain cells, bar none." His reasoning reveals why venture capitalists are betting so heavily on robots: the market combines the shipment scale of smartphones with the unit price of passenger vehicles. If the industry matures as expected, with annual shipments reaching 1 billion units at an average price of 30,000 USD (roughly 200,000 RMB), the total addressable market would rival some of the largest commercial sectors in human history.

The secondary market has responded with enthusiasm. When the first one or two companies in a major AI track go public, they often enjoy what Zhou calls a "scarcity dividend," where stock prices and valuations deviate from conventional logic. This dynamic has created a rush of new entrants; Qiming's team estimates that 370 new embodied intelligence companies were established domestically in just the past two years, with two to three new business plans arriving each week.

However, Zhou cautioned that the field faces a critical challenge: many companies are becoming difficult to distinguish. Founding teams cluster into a few categories, including university professors, senior executives from autonomous driving companies, and AI model researchers. Technical approaches converge around two main routes: Vision Language Models (VLMs) and world models, with implementation scenarios overlapping in industrial logistics, commercial services, and bionic robots. Without key breakthroughs and technical convergence, Zhou warned, the industry risks producing only "dispensable demo projects" rather than scaled commercial solutions.

Zhou

Why Does China Have an Unfair Advantage in Robotics?

Zhou identified three structural advantages that position China ahead of the United States in embodied intelligence development. First, data reserves are more abundant in China; several leading U.S. technology giants are purchasing data from Chinese enterprises, a sign of insufficient domestic supply. A data platform company invested by Qiming, established only four months prior to the July 2026 announcement, already had orders totaling 1 billion RMB.

Second, China possesses a vast ecosystem of manufacturing and implementation scenarios. Companies like CATL and BYD operate large physical factories capable of collaborating on research and development, offering embodied intelligence startups access to real-world industrial environments. By contrast, leading U.S. companies such as Figure AI have had to travel to Europe to partner with manufacturers like BMW.

Third, and perhaps most critically, China controls the supply chain for robot hardware. A humanoid robot contains approximately 1,200 components, and more than 90% of the global supply chain is concentrated in the Yangtze River Delta and Pearl River Delta regions of China. This proximity allows domestic embodied intelligence companies to rapidly iterate on both robot bodies and AI models simultaneously. When engineers discover a mismatch between model algorithms and hardware execution, they can contact suppliers and adjust within two weeks, a speed advantage that U.S. competitors cannot match.

How to Navigate the Crowded AI Investment Landscape

• Track Industry Convergence: Monitor whether technical routes in embodied intelligence are converging toward unified standards, as fragmentation signals higher risk of market failure and valuation inversions between primary and secondary markets.
• Evaluate Supply Chain Proximity: Assess whether AI robotics companies have direct access to hardware suppliers and manufacturing partners, as this determines iteration speed and competitive advantage in bringing products to market.
• Examine Data Moats: Investigate whether companies have proprietary data reserves or partnerships that provide exclusive training material, since data scarcity is now a limiting factor even for leading U.S. technology giants.

Zhou also addressed the recent hype around world models, a technical approach that has attracted approximately 30 new startups. He stated frankly that world models are not a fundamentally new track; newly emerged startups pursuing this approach have no essential difference in commercial implementation compared with companies that previously adhered to VLA technical routes. "In other words, the world model has just been hyped into a popular concept in the primary market," he noted.

What Is Kling AI's Role in Qiming's Broader Portfolio?

Kling AI, a video generation tool, represents Qiming's investment in the creative AI applications layer. The platform is described as a sophisticated tool for generating images and videos, tailored for creators who want to craft imaginative content using advanced generative AI techniques. Its distinctive feature is the dual binding of visual identity and vocal tone, which facilitates a smooth approach to multimedia storytelling.

Kling AI sits within Qiming's systematic layout across the entire AI industry stack. The firm's portfolio spans from chips carrying computing power and foundation large models driving innovation, to robots and vertical scenario applications, extending into cutting-edge integrated fields such as life sciences. This comprehensive approach began as early as 2013, when the firm invested in early-stage companies like iFlytek, Megvii, UBTECH, and WeRide. Before the emergence of GPT-3 in 2020 and the global popularity of ChatGPT at the end of 2022, Qiming had already positioned itself at key nodes in the AI industry chain through investments in companies like Zhipu, StepStar, and others.

The firm's track record demonstrates the value of patient capital and deep industry expertise. Zhou Zhifeng has witnessed the concentrated IPOs of Biren Technology, Zhipu, Axera, and Yunyinggu, validating the firm's long-term cultivation of emerging AI sectors. As the AI market matures and valuations stabilize, Qiming's emphasis on comprehensive information collection and continuous tracking of the industry landscape positions it to identify winners before the broader market recognizes them.

"The more frenetic the era, the more we need to learn from historical lessons; the more noisy the era, the more important philosophical thinking becomes," remarked Zhou Zhifeng, Managing Partner of Qiming Venture Partners.

Zhou Zhifeng, Managing Partner at Qiming Venture Partners

For investors, entrepreneurs, and AI observers, Qiming's strategy offers a clear signal: the next wave of AI value creation will not come from foundation models alone, but from the intersection of hardware, data, and implementation scenarios. Companies that can rapidly iterate on both algorithms and physical systems, backed by deep supply chain integration, will likely emerge as winners in embodied intelligence. Meanwhile, video generation tools like Kling AI represent the creative applications layer that will drive consumer adoption and revenue growth across the broader AI ecosystem.