According to Smart Emergence, the embodied intelligence company StarSea Map has recently successfully completed its A4 and A5 strategic financing rounds, with a total funding exceeding 100 million US dollars. The lead investors in this round are Meituan Longzhu and Today's Capital, while existing shareholders Mihoyo and Wuxi Investment Group also participated. This financing marks Meituan's continued investment and layout in the field of embodied intelligence.
According to internal sources, StarSea Map explained that it unified all financing this year under the name "Series A" based on its judgment of the current stage of industry development. Industry experts generally believe that embodied intelligence is still in the infrastructure construction phase and has not yet entered the stage of large-scale application. Therefore, the company chose to raise funds in stages to better adapt to market demands.
In recent years, Meituan has significantly increased its investment frequency in the field of embodied intelligence. In May 2025, Meituan's venture capital led the Series A financing of several hundred million yuan for Self-Variable Robotics, and StarSea Map's Series A financing is another important investment in this field. Meituan has already made investments in multiple embodied intelligence companies, including Unitree Technology and Galaxy General.
Since the start of its Series A financing in 2025, StarSea Map has raised nearly 1.5 billion RMB in total. The latest valuation has grown more than three times compared to the beginning of the year. These funds are mainly used to enhance the company's "full machine + intelligence" full-stack self-research capabilities, especially in improving the end-to-end model capabilities of VLA (Visual-Language-Action), expanding production scale, and developing joint modules.
The core team of StarSea Map consists of multiple AI and embodied intelligence algorithm scientists. CEO Gao Jiyang has rich experience in the field of autonomous driving and previously served as a senior engineer at Waymo. The team's research covers core technologies such as reinforcement learning, imitation learning, and whole-body motion control, ensuring the company's competitiveness in the industry.
In addition, StarSea Map has introduced a "fast-slow" dual system model architecture for embodied intelligence — EFM-1, which combines visual language models and action models, aiming to achieve closed-loop decision-making from perception to control. The company is also committed to opening up these basic models and data to promote the construction of a global developer ecosystem.
As the industry's attention to embodied intelligence continues to deepen, StarSea Map's rapid development and Meituan's investment strategy will further drive innovation and application in this field. It is expected that future breakthroughs will be achieved in multiple scenarios, especially in intelligent manufacturing and service industries.