Author: brianbreslow, Hypersphere Ventures
Translated by: Tim, PANews
Executive Summary
- Humanoid general-purpose robots are rapidly transitioning from science fiction to reality. Three converging factors are actively driving the next major platform iteration in computing: continuously declining hardware costs, sustained capital investment, and technological breakthroughs in motion flexibility and operational capabilities.
- Despite increasing computational power and hardware commoditization offering low-cost advantages for robotics engineering, the industry remains constrained by training data bottlenecks.
- Reborn is one of the few projects utilizing decentralized Physical AI (DePAI) to crowdsource high-precision motion and synthetic data, and build robot foundation models, placing it in a unique advantageous position for advancing humanoid robot deployment. The project is led by a founding team with profound technical expertise, whose members have academic research experiences and professorial backgrounds from UC Berkeley, Cornell University, Harvard University, and Apple, demonstrating both exceptional academic standards and real-world engineering execution capabilities.
Humanoid Robots: From Science Fiction to Cutting-Edge Applications
Robotics commercialization is not a new concept. Well-known consumer robots like the iRobot Roomba vacuum cleaner from 2002 or recent home devices like the Kasa pet camera are single-function devices. With AI development, robots are evolving from single-function machines to multifunctional forms designed to adapt to tasks in open environments.
Humanoid robots will gradually upgrade from basic tasks like cleaning and cooking to eventually handle complex work such as reception services, firefighting, and even surgical operations within the next 5 to 15 years.
Recent developments are transforming humanoid robots from science fiction into reality.
- Market Dynamics: Over 100 enterprises are positioning themselves in humanoid robotics (such as Tesla, Unitree, Figure AI, Clone, Agile, etc.).
- Hardware technology has successfully crossed the uncanny valley: The new generation of humanoid robots demonstrates fluid, natural movements, enabling human-like interactions in real environments. For instance, Unitree H1 can walk at speeds up to 3.3 meters per second, far exceeding the human average of 1.4 meters per second.
(Note: The Uncanny Valley theory is a psychological concept describing human emotional responses to non-human entities like robots, dolls, and virtual representations.)
- New Cost Paradigm for Humanoid Robots: Expected to fall below US labor wage levels by 2032.
Development Bottleneck: Real-World Training Data
Despite clear positive factors in the humanoid robot field, low data quality and scarcity continue to impede large-scale deployment.
Other AI entity technologies, such as autonomous driving, have essentially solved data problems through existing vehicle-mounted cameras and sensors. Using Tesla and Waymo as examples, these fleets can generate billions of miles of real road driving data. During Waymo's development stage, human supervisors were present in the passenger seat for real-time training.
However, consumers are unlikely to accept a "robot babysitter". Robots must demonstrate high performance out of the box, making pre-deployment data collection crucial. All training must be completed before commercial production, with data scale and quality remaining persistent challenges.
Although each training mode has its own scale unit (e.g., tokens for large language models, video-text pairs for image generators, motion segments for robotics), the comparison below clearly reveals the magnitude gap in robotics data availability:
- GPT-4's training data exceeds 150 trillion text tokens.
- Midjourney and Sora utilize billions of labeled video-text pairs.
- In contrast, the largest robot dataset contains only about 2.4 million interaction records.
This gap explains why robotics has not yet achieved a true foundation model like large language models, with the key issue being an incomplete data foundation.
Traditional data collection methods cannot meet the scalability requirements for humanoid robot training data. Existing methods include:
- Simulation: Low-cost but lacking real boundary scenarios (simulation-reality gap)
- Internet videos: Cannot provide the proprioception and force feedback environments necessary for robot learning
- Real-world data: Accurate but requiring remote control and human-loop operations, resulting in high costs (over $40,000 per robot) and limited scalability.
Training models in virtual environments is low-cost and scalable, but these models often struggle when deployed in the real world. This problem is known as the Sim2Real gap.
For example, a robot trained in a simulated environment might easily grasp objects with perfect lighting and smooth surfaces but become helpless when facing cluttered environments, textured surfaces, or the various unexpected situations humans encounter in the real world.
Reborn offers an economical and efficient method to crowdsource real-world data, helping to strengthen robot training and bridge the Sim2Real gap.
Reborn: The Full-Stack Vision of Decentralized Physical AI
Reborn is constructing a vertically integrated software and data platform for embodied intelligent robot applications. While the company's core goal is to address data bottlenecks in humanoid robotics, its vision extends far beyond. By combining self-developed hardware, multi-modal simulation infrastructure, and foundation models, Reborn will become a full-stack driver for embodied intelligence.
The Reborn platform begins with a proprietary consumer-grade motion capture device called "ReboCap", building a rapidly expanding augmented and virtual reality gaming ecosystem. Users provide high-fidelity motion data in exchange for network incentive rewards, driving platform development. Currently, Reborn has sold over 5,000 ReboCap devices, with 160,000 monthly active users and a clear growth path to exceed 2 million users by year-end.
Reborn's data collection benefits significantly outperform other solutions.
Notably, this growth is entirely organic: users are attracted by the game's entertainment value, and streamers use ReboCap to drive real-time body tracking for digital avatars. This spontaneously formed virtuous cycle achieves scalable, low-cost, high-fidelity data production, making the Reborn dataset a training resource sought after by top robotics companies.
The second layer of the ReBorn software stack is Roboverse: a multi-modal data platform that unifies fragmented simulation environments. The current simulation field is highly fragmented, with tools like Mujoco and NVIDIA Isaac Lab operating independently, each with advantages but lacking interoperability. This fragmentation delays research and development and exacerbates the simulation-reality gap. Roboverse standardizes multi-simulator integration, creating a shared virtual infrastructure for robot model development and evaluation. This integration supports consistent benchmarking and significantly enhances system scalability and generalization.
Roboverse enables seamless collaboration. The former massively collects real-world data, while the latter builds simulation environments to drive model training, jointly demonstrating the true power of Reborn's distributed physical intelligence network. The platform is creating a physical AI developer ecosystem that extends beyond mere data acquisition, with functions now reaching actual model deployment and commercial licensing.
Reborn Foundation Model
Perhaps the most critical component in Reborn's technology stack is the Reborn Foundation Model (RFM). As one of the first robot foundation models, it is being developed as the core system for emerging physical AI infrastructure. Its positioning is similar to traditional large language foundation models like OpenAI's GPT-4 or Meta's Llama, but oriented towards the robotics domain.
Reborn Technology Stack
The three core components of the Reborn technology stack (ReboCap data platform, Roboverse simulation system, and RFM model authorization mechanism) together build a solid vertical integration moat. By combining crowdsourced motion data with a powerful simulation system and model authorization system, Reborn can train a foundational model with cross-scenario generalization capabilities. This model can support diverse robot applications in industrial, consumer, and research fields, achieving universal deployment under massive and diverse data.
Reborn is actively promoting the commercialization of its technology, launching paid pilot projects with Galbot and Noematrix, and establishing strategic partnerships with Unitree, Booster Robotics, Swiss Mile, and Agile Robots. The humanoid robot market in China is experiencing rapid growth, accounting for approximately 32.7% of the global market. Notably, Unitree occupies over 60% of the global quadruped robot market and is one of the six Chinese manufacturers planning to produce over 1,000 humanoid robots by 2025.
The Role of Cryptocurrency Technology in Physical AI Technology Stack
Cryptographic technology is building a complete vertical stack for physical world AI.
Reborn is a Leading Embodied AI Cryptocurrency Project
Although these projects belong to different layers of the physical AI stack, they share one common point: they are 100% DePAI projects. DePAI creates an open, composable, and permissionless expansion mechanism through token incentives throughout the technology stack, and it is this innovation that makes the decentralized development of physical AI a reality.
Reborn has not yet issued a token, making its organic business growth even more commendable. When the token incentive mechanism is officially launched, network participation will accelerate as a key link in the DePAI flywheel effect: users can receive project incentives by purchasing Reborn hardware devices (ReboCap collectors), and robot R&D companies will pay contribution rewards to ReboCap holders. This dual incentive will drive more people to purchase and use ReboCap devices. Meanwhile, the project will dynamically incentivize high-value customized behavior data collection, thereby more effectively bridging the technical gap between simulation and real-world applications (Sim2Real).
Reborn's DePAI Flywheel in Operation
The "ChatGPT moment" in robotics will not be triggered by robot companies themselves, as hardware deployment is far more complex than software. The explosive growth of robotic technology is naturally limited by cost, hardware availability, and deployment complexity, obstacles that do not exist in purely digital software like ChatGPT.
The turning point for humanoid robots is not how impressive the prototype is, but whether the cost drops to a range affordable by the masses, similar to the popularization of smartphones or computers in the past. When costs decrease, hardware will become merely an entry ticket, and the real competitive advantage will lie in data and models: specifically, the scale, quality, and diversity of motion intelligence used to train machines.
Conclusion
The robot platform revolution is unstoppable, but like all platforms, its large-scale development cannot be separated from data support. Reborn, as a high-leverage bet, firmly believes that cryptographic technology can fill the most critical gap in the AI robot technology stack: its robot data solution, DePAI, is cost-effective, highly scalable, and modular. As robotic technology becomes the next frontier of AI, Reborn is transforming the general public into "miners" of motion data. Just as large language models require text markup support, humanoid robots need massive motion sequence training. Through Reborn, we will break through the final bottleneck and realize the leap of humanoid robots from science fiction to reality.
