China’s embodied-AI investment cycle is increasingly reaching beyond robot brands and into the parts that determine how machines move. Nanjing Encos Intelligent Technology, which uses the ENCOS brand and is also known as Inx in some reports, completed a Series B financing round of more than 300 million yuan on September 14. The round puts fresh capital behind integrated robot joint modules, a category that sits between a robot’s high-level software and its mechanical movement.
36Kr reported the financing, while Sina’s coverage identified CITIC Goldstone and NICE Group among the new participants. The reports also named Sinorich Investment, Nanjing Communications Control, Warui Venture Capital, and Suzhou Venture Capital. Existing backers that joined the round included Puhua Capital, Jinqiu Fund, Huakong Fund, Shenzhen Venture Capital, and Fosun Fortune. The exact round size was not disclosed beyond “more than 300 million yuan.”
ENCOS makes integrated joint modules and related components for embodied-intelligence robots. That focus makes the financing a different kind of physical-AI story from the robot-factory announcements that dominate headlines. A humanoid’s visible body may drive public attention, but joint systems determine how a robot converts a control command into mechanical movement and how much hardware integration remains for the robot maker.
Why Robot Joints Are Becoming an Investment Category
An embodied system requires models, sensors, software, mechanical structures, and actuators to work as one product. ENCOS is concentrating on the actuator side: the integrated systems that help a robotic arm, leg, or other moving assembly produce controlled motion. The company’s financing is therefore a bet that component specialization can scale alongside demand for humanoid and other embodied machines.
This is not a trivial layer of the supply chain. A robotics company may differentiate through perception or a foundation model, but it still needs hardware that can execute a motion repeatedly within power, heat, size, and cost constraints. A joint module combines several such requirements into a unit that can be selected and integrated by a robot maker. The appeal of an established module is that it may reduce the engineering work needed to design every actuator from the ground up.
EastFrontier has previously tracked the industry through large public-facing milestones, including China’s first automated humanoid robot factory and a record early financing for TARS in China’s humanoid-robotics sector. ENCOS’s round points at a parallel question: which suppliers will provide the movement systems for the expanding number of robot programs?
The answer will not be determined by financing alone. Component suppliers need customers that can move from prototypes to deployments, and robot makers need parts that meet their own requirements. Still, a round above 300 million yuan gives ENCOS more runway to pursue manufacturing, product development, and customer support in a market where hardware iteration can be capital intensive.
The Second-Generation Module and a Platform Ambition
Sina reported that ENCOS released a second-generation quick-release joint module for humanoid robots in June 2026. The company has also said it plans to launch an online hardware-selection platform by the end of the year. Those two details describe a strategy that is broader than supplying a single fixed component.
A quick-release design may matter because it can make a robot’s mechanical configuration easier to assemble, test, replace, or alter. ENCOS calls its module an industry first, a claim that should be understood as the company’s own characterization rather than an independently established ranking. The company has also described performance improvements for its joint modules. Those technical claims require the same caution: they are product statements, not third-party benchmark results.
The selection-platform plan is potentially more important than it sounds. A catalog and configuration layer could give customers a way to compare modules for different robotic tasks, then build systems around standardized options. ENCOS has not released full details of the platform, its customer access model, or its pricing. But the stated plan indicates that it wants to become easier for robot developers to evaluate and specify hardware, not merely purchase it through one-off relationships.
That ambition mirrors a broader pattern in China’s hardware ecosystem: the winners are often not only the firms with a strong component, but the firms that make component selection, supply, support, and iteration easier for downstream customers.
Capital Will Not Eliminate the Hardware Test
The financing creates an opportunity, not a conclusion. ENCOS must turn new capital into components that meet robot makers’ real needs and into supply capacity that can support actual orders. In embodied AI, a component’s value is tested through integration: whether it works with control software, fits a mechanical design, survives repeated use, and can be repaired or replaced without disrupting an entire system.
The company’s investor list also suggests that the round attracted a mix of strategic and venture capital. That mix may help a hardware supplier build relationships across manufacturing, finance, and robotics ecosystems. It does not guarantee product adoption. The number of humanoid projects in China has grown rapidly, but the long-term component market will depend on how many of those projects reach sustained deployment rather than limited demonstrations.
ENCOS’s more than 300 million yuan Series B is therefore notable for what it funds at the base of the stack. While models decide what a robot should attempt, joint systems determine whether a physical machine can make the commanded movement. As China’s embodied-AI sector shifts from prototype displays toward more demanding use cases, that distinction will become increasingly visible.