Jijia Vision, a Beijing-based artificial intelligence startup operating under the name GigaAI, is in discussions to hold an initial public offering on the Hong Kong stock exchange as early as this year. The company, which focuses on developing “world models” for physical AI applications, is currently closing a new funding round that will value it at $3 billion. If successful, the listing would make GigaAI the first global world-model startup to go public, highlighting the rapid commercialization of embodied AI technologies in China.
Founder and CEO Huang Guan, a former researcher at Microsoft and Horizon Robotics, confirmed the IPO plans during an interview with Bloomberg on the sidelines of the World Artificial Intelligence Conference in Shanghai. The $3 billion valuation would rank GigaAI as the most valuable startup in China focusing on world models, AI systems designed to understand and predict the physical dynamics of the real world, which are crucial for the development of advanced robotics and autonomous vehicles.
Capitalizing on China’s Industrial Base
GigaAI’s rapid ascent underscores the strategic advantage Chinese AI startups hold in physical and industrial applications. While Western world model startups, including those backed by AI pioneers Fei-Fei Li and Yann LeCun, have attracted significant investment, GigaAI is leveraging China’s extensive manufacturing ecosystem to accelerate deployment. “We have diverse commercial and industrial application scenarios that aren’t available in the US or elsewhere,” Huang stated, emphasizing that China’s mature industrial base is far ahead of the rest of the world.
The startup’s product portfolio includes models specifically tailored for autonomous driving systems and humanoid robots. By focusing on practical, factory-floor applications, GigaAI has adopted a highly pragmatic approach to AI development. “We move very quickly from research to engineering to product to business, creating value for our customers,” Huang explained. This focus on rapid commercialization aligns with Beijing’s national priorities, which have designated embodied AI and humanoid robotics as key strategic sectors for future economic growth.
The Broader AI IPO Wave
GigaAI’s planned listing is part of a broader wave of Chinese artificial intelligence companies preparing to tap the public markets, a trend that has seen Zhipu AI and MiniMax post their first earnings as public companies with starkly different outcomes. As the competition with US tech giants intensifies, Chinese startups are racing to secure the capital necessary to fund expensive model training and acquire scarce computing resources. The push for public listings is also driven by the need to provide liquidity for early investors and attract top-tier talent through stock options.
The impending IPOs of companies like GigaAI, DeepSeek, and Moonshot AI signal a maturation of the Chinese AI ecosystem. While American rivals such as Anthropic and OpenAI are also reportedly considering Wall Street debuts, the flurry of activity in Hong Kong and Shanghai suggests that Chinese firms are moving aggressively to solidify their financial foundations. As world models transition from research labs to real-world applications, access to public capital will be a critical factor in determining which companies ultimately dominate the embodied AI sector.
What Makes a World Model Different
World models are distinct from conventional large language models in that they are designed to learn and simulate the physical laws governing the real world. Rather than simply predicting the next word in a sequence, a world model predicts the next state of a physical environment, enabling robots and autonomous vehicles to plan actions, anticipate consequences, and adapt to unexpected situations. This capability is fundamental to the development of truly autonomous physical AI systems.
GigaAI’s competitive advantage lies in its ability to train these models on rich, diverse data drawn from China’s industrial environments. The company has deployed its technology across automotive, logistics, and manufacturing clients, accumulating proprietary real-world data that is difficult for competitors to replicate. Huang Guan has noted that China’s sheer scale of industrial activity, from its vast network of factories to its rapidly expanding fleet of autonomous vehicles, provides a data advantage that Western startups simply cannot match. As the race to build capable physical AI systems intensifies, this data moat may prove to be GigaAI’s most durable competitive asset.
For investors, the GigaAI IPO represents an opportunity to gain exposure to one of the most compelling long-term themes in technology: the convergence of AI with the physical world. As humanoid robots, autonomous vehicles, and intelligent manufacturing systems become increasingly prevalent, the companies that can provide the underlying intelligence layer stand to capture enormous value.
GigaAI’s planned listing, if successful, is likely to serve as a bellwether for the entire world model sector, potentially opening the floodgates for a wave of similar listings from Chinese physical AI startups in the months and years ahead. The company’s decision to pursue a Hong Kong listing rather than a domestic STAR Market debut also reflects a deliberate strategy to attract international institutional capital and build a global investor base, signaling confidence that the world model thesis resonates beyond China’s borders.