BridgeDP raises nine-figure round to accelerate its general-purpose robot operating system
The round was led by China Mobile Chain Fund, with existing investors Fosun Capital and Shenzhen Capital Group participating

BridgeDP Robotics recently announced the completion of a Pre-A+++ funding round. The round was led by China Mobile Chain Fund, with existing investors Fosun Capital and Shenzhen Capital Group participating. The financing reached the hundreds of millions of yuan scale.
With this round, BridgeDP further clarified its long-term strategic direction: building a 'general-purpose robot operating system' on the foundation of motion control. This marks our transition from a motion control solution provider to a builder of infrastructure and ecosystem for the general-purpose robotics industry.
Within six months, BridgeDP has received investments from industrial capital and state-owned funds. This is both a recognition of our past achievements and an affirmation of and trust in our new positioning and direction.
This round of funding will be used to accelerate the implementation of this strategy and advance our product iteration and algorithm research to the next stage.
RoboCraft AI: a platform for developing general-purpose robot motion capabilities
We released RoboCraft AI in June this year. It brings together BridgeDP's experience from the development and deployment of more than 50 different robot models, distilling underlying motion control technologies, model algorithms, and Sim2Real engineering expertise into standardized software development processes.
We hope that developers can use RoboCraft AI to quickly build first-class motion capabilities for robots.
On one hand, for multiple mainstream and mature robot models, the platform supports rapid development and on-device deployment of application scenarios.
RoboCraft AI is compatible with models such as Unitree G1, AGIBOT X2, EngineAI PM01, and Booster T1. For lessors, integrators, and content teams that have purchased these models, we provide an out-of-the-box application development toolchain that consolidates motion customization, application orchestration, and on-device deployment into a single workbench. Specifically, users can:
- Use the platform's high-quality asset library of walking, running, dancing, and other motions;
- Add custom motions via 'video extraction' or 'upload BVH files';
- Orchestrate multiple motions into applications with simple drag-and-drop on the timeline, and save and manage them over time;
- Deploy applications to real robots with one click and control multiple robots simultaneously.
During a one-month closed beta, RoboCraft AI has established connections and cooperation with dozens of robot rental and performance companies nationwide. Several customers have independently completed the deployment process and significantly improved development efficiency for commercial performance scenarios using the platform.
On the other hand, the platform also supports integration and adaptation of proprietary robot models from OEMs.
For companies developing their own robots, RoboCraft AI provides a standardized onboarding process that condenses previously expensive, expert-dependent work into reusable automated platform features, significantly shortening the cycle for deploying motion capabilities on new robot models.
RoboCraft AI has developed functions from AI validation and auto-repair of URDF files, to hardware safety checks, to automatic generation of training parameters. It attempts to resolve hardware risks in Sim2Real engineering upfront, improving the success rate and efficiency of subsequent deployments.
We have reached deep cooperation with multiple robot OEMs and joint module manufacturers, completing 0-to-1 integration and adaptation for their new models. We have helped partners obtain basic motion capabilities and room for secondary development comparable to mature robot platforms.

RoboCraft AI is now fully open for registration. Visit robocraftai.cn to access the high-quality built-in motion library and the complete application development toolchain. We welcome industry partners to try it and share feedback!
A general-purpose motion control model: a powerful foundation for motion capabilities
The cross-platform deployment capability and motion generalization demonstrated by RoboCraft AI stem from BridgeDP's general-purpose motion control model.
Our algorithm team integrated motion data and deployment experience accumulated from serving different robots into the development of a general-purpose motion tracker. Based on our self-developed general-purpose motion tracking model architecture and trained on thousands of hours of self-collected full-body motion data across different robot platforms, it can now reliably support high-frequency motions such as walking, running, jumping, dancing, and martial arts. It also supports many dance choreographies that had not been individually adapted before, demonstrating strong motion generalization.
As the production capacity of BridgeDP's cross-platform full-body motion data factory increases, the model's motion coverage, complex motion generalization, and deployment stability will continue to improve.
Moreover, the general-purpose motion tracker is not just a single motion model; it can also take input from upper-layer motion planners. We have made it compatible with the outputs of both explicit and implicit walking planners, converting upper-layer planning results into executable and deployable full-body motion control capabilities.
In the future, based on this foundation, our interface capabilities will gradually expand to more complex tasks such as navigation and manipulation. This will extend the motion control model from mere 'motion execution' to the core basic motion capability layer of the 'general-purpose robot operating system' that BridgeDP aims to build.
Cross-platform full-body motion data factory: the key to continuous capability evolution
In early June this year, our self-built cross-platform full-body motion data factory was put into use. It has already produced the first batch of thousands of hours of effective training data that passed full-process quality inspection. The production line has moved from validating the model to ramping up capacity.
In data planning, we adopt a dual-track production model of 'proactive coverage + model capability feedback':
- For proactively designed data collection, a dedicated motion design team works with an AI-native management platform to continuously expand the motion library across capability dimensions such as locomotion, posture transitions, full-body interaction, and fall recovery, ensuring data diversity and effectiveness;
- For data collection driven by model capability requirements, feedback such as training failures, cross-platform transfer degradation, dynamics regressions, and low-yield samples is used to optimize the next round of collection and retargeting, so that data production serves the real gaps in model iteration.
At the same time, the data factory is equipped with self-developed cross-platform motion retargeting algorithms that can quickly adapt and distribute collected data to different robot models. This enables efficient reuse of high-quality motion data across multiple robot platforms, significantly improving data production and application efficiency, and further accelerating the training, validation, and capability evolution of the cross-platform motion control model.

Once again, we thank all investment institutions and industry partners for their long-term recognition and trust in BridgeDP's capabilities in motion control.
Going forward, while continuing to deepen our core business strengths, we will increase investment in the data factory, algorithm research, and product iteration to lay a solid foundation for building the 'general-purpose robot operating system'.
We hope to provide the industry with an infrastructure for general-purpose robot application development, allowing more people to participate in the future ecosystem of robotics. We believe this is a good path to improving industry efficiency and driving industrial innovation.
As BridgeDP founder Shang Yangxing said: 'BridgeDP's mission is to let intelligence truly enter and change the physical world. To achieve this, we cannot rely on just a few companies or a few genius engineers. We want more developers, hardware teams, scenario experts, and entrepreneurs to participate in building physical AI.'





界面新闻
深视新闻
深圳特区报
财新
36氪
钛媒体
中国新闻周刊
蓝鲸新闻
Z Finance
投资家网
雷峰网
亿欧
环球网
南方日报
极客公园



