The global artificial intelligence race is shifting its gaze from the digital realm of text and images to the tangible world of atoms and motion. Leading this charge in Asia is 51World, a digital twin pioneer that recently debuted on the Hong Kong Stock Exchange. At its 2025 performance briefing, Chairman Li Yi articulated a hierarchy of AI value that places the burgeoning 'Physical AI' sector at the top, projecting it as a $10 trillion market—dwarfing both Generative AI and Agentic AI.
To capture this massive opportunity, 51World is pivoting its business model from traditional software licensing to what Li calls a 'Physical AI Factory.' This transition moves the company away from simply selling simulation tools toward providing an integrated service that combines massive computing power, specialized software, and synthetic data. Under this new framework, clients pay based on usage and computational results, a shift aimed at building recurring, stable service revenue while lowering the entry barrier for industrial partners.
Nowhere is this strategy more evident than in the autonomous driving sector. 51World currently commands a 53.5% market share in China’s end-to-end high-level simulation market. The company recently sacrificed short-term profit margins—a strategic 'access battle' as Li describes it—to secure long-term dominance. By embedding its 51Sim platform into the development pipelines of 60% of China’s top carmakers, the firm is positioning itself as the indispensable infrastructure for the next decade of mobility.
However, the company’s ambitions extend far beyond the asphalt. The 'Physical AI Factory' is designed to be a cross-industry engine, providing the virtual training grounds for embodied intelligence, such as humanoid robots, and the rapidly growing 'low-altitude economy' involving drones and eVTOL aircraft. As these machines require millions of hours of simulated training before they can safely interact with the real world, 51World’s ability to provide high-fidelity synthetic environments is becoming a critical bottleneck for the entire industry.
By 2030, the company anticipates its digital twin business will provide a stable foundation while its simulation and 'Physical AI' segments enter an explosive growth phase. This roadmap suggests a future where the distinction between the physical and digital worlds blurs, as AI agents are perfected in 'factories' of pure data before they are ever deployed into the real-world environments of cities, factories, and homes.