36Kr Reports on Investment Trends in ‘Embodied Intelligence’

On October 22, 2026, 36Kr reported on “Embodied Intelligence: Hot Money Abandons Traditional Ties to Chase Next-Gen AI Investment Opportunities.” The source is an interview, column or opinion piece, with the statements and arguments at its core; this summary notes only that the report was published.

Because Embodied Intelligence presupposes machines or devices that operate in the real world, rather than software alone, the factors informing investment decisions are thought to differ from those for conventional AI. 36Kr’s coverage of these capital flows may offer insight into how investors assess not only research and development but also the potential for mass production and deployment in the field.

36Kr addressed the argument that investment interest in Embodied Intelligence is moving away from traditional ties and toward next-generation AI opportunities. Rather than retracing the original’s statements and claims, we would first note that this topic is being reported as a shift in the direction of capital, not simply as a development in the technology itself. AI with a physical presence involves connected stages, including computing infrastructure, robots, sensors, data collection and deployment in the field. Investors therefore need to determine which layers are attracting funding, rather than focus only on the performance of individual models. No previous reports are cited as related coverage by this newspaper, so it is not possible to trace a direct connection here. At the same time, based on publicly available information, Embodied Intelligence is seen as a field that presupposes hardware, control systems and a cycle of data from real-world operations, rather than software alone, as in generative AI. As a result, even when funding flows into the sector, evaluation criteria are likely to shift from the appeal of a model to the maturity of prototypes, the potential for mass production, safety and operational track records at deployment sites. The original’s detailed arguments should be checked directly, including the context of the statements. Three points warrant attention going forward. First, which technology layers are attracting investment. Second, whether evaluation of Embodied Intelligence is moving from the prototype stage to mass production and deployment. Third, how the collection and use of real-world data for training are being treated as prerequisites for monetization.

In Japan, many companies are seen as having strengths in physical components such as robots, control systems, sensors and manufacturing equipment, while structural gaps can emerge in the accumulation of real-world data and the scale of fundraising. If investment in Embodied Intelligence continues, a key issue for Japanese companies will be how to build systems that connect deployment with learning, rather than focusing only on individual components.