Singapore-based Ropedia, a startup developing advanced data infrastructure for physical AI, has announced the successful completion of a $30 million Pre-A funding round. The investment will enable the company to rapidly expand its multimodal data collection platform, accelerate AI research, and strengthen deployments across Southeast Asia and North America.
The funding attracted venture investors with extensive expertise in artificial intelligence, enterprise technology, and infrastructure, alongside strategic partners from robotics, mobility, and enterprise deployment sectors. The latest financing reinforces growing investor confidence in the future of embodied AI and robotics.
Funding to Drive the Next Era of Physical AI
Unlike traditional AI models trained primarily on internet text, physical AI systems require exposure to real-world human interactions. Robots must understand how humans perform physical tasks rather than simply observing images or videos.
Explaining the company’s vision, Zhaoxi Chen, CEO and Co-founder of Ropedia, said:
“A robot can’t play baseball simply by watching a video, just as people can’t learn to ride a bicycle by reading a book. Physical AI requires real-world experience. Human interaction data will become the foundation for training the next generation of intelligent robots that will eventually assist in factories, businesses, and homes.”
According to Chen, physical AI represents the next major shift in artificial intelligence, moving beyond language understanding toward machines capable of performing meaningful real-world tasks.
Building End-to-End Data Infrastructure Instead of Simple Data Labeling
Ropedia differentiates itself from conventional data annotation companies by creating a complete end-to-end infrastructure for physical AI.
Rather than labeling existing datasets, the company captures, synchronizes, structures, processes, and continuously refines entirely new multimodal datasets specifically designed for robotics training.
Its integrated platform combines:
- Proprietary wearable hardware
- Multimodal data capture
- Automated synchronization
- Quality assurance systems
- AI-ready structured datasets
- Model fine-tuning infrastructure
This vertically integrated approach enables AI companies to receive production-ready datasets rather than fragmented data collection services.
HOMIE: The Wearable Device Powering Physical AI
At the center of Ropedia’s technology ecosystem is HOMIE, a wearable head-mounted device engineered to capture comprehensive first-person human experiences.
The device simultaneously records:
- First-person video
- Audio
- Depth information
- Hand tracking
- Eye gaze
- Full-body motion
- Camera positioning
Every sensor stream is synchronized with precise timestamps, enabling AI models to accurately correlate perception with physical movement.
This synchronized multimodal data forms the foundation for robotics systems learning complex human behaviors.
Unlike teleoperation-based data collection, which depends on expensive robot fleets, HOMIE allows scalable human-based data collection simply by deploying additional wearable devices.
Closed-Loop Data Pipeline Delivers High-Quality AI Training Data
Ropedia has built what it calls a closed-loop data pipeline, integrating every stage of physical AI dataset generation.
The pipeline includes:
- Real-world multimodal data collection
- Automatic synchronization
- Quality assurance
- Data refinement
- Model-aligned optimization
The company commercializes this infrastructure through three primary offerings:
- Dataset licensing
- Selective HOMIE hardware access
- Research collaborations with AI and robotics companies
This flexible business model enables organizations developing robotics and embodied AI systems to integrate high-quality training data directly into their development workflows.
Xperience-10M: One of the World’s Largest Human Interaction Datasets
Supporting the platform is Xperience-10M, one of the industry’s largest datasets dedicated to physical AI.
The dataset currently contains:
- More than 10 million human interaction episodes
- Over 10,000 hours of multimodal recordings
- Billions of synchronized frames across:
- Video
- Depth sensing
- Motion capture
- Inertial sensor data
Each additional HOMIE deployment continuously expands the dataset by introducing new environments, behaviors, and interaction patterns, increasing its long-term value for robotics developers.
Data Collection Costs Reduced by Up to 50 Times
One of Ropedia’s most significant competitive advantages is operational efficiency.
According to the company, its wearable-based infrastructure reduces data collection costs by up to 50 times compared to traditional robotics data acquisition methods.
HOMIE has already entered mass production, allowing Ropedia to rapidly scale deployments while maintaining lower infrastructure costs than teleoperation-based systems.
The startup has already provided services to more than a dozen North American companies specializing in:
- Embodied AI
- Robotics
- Spatial intelligence
- Foundation AI models
Breakdown of the $30 Million Funding Round
The total $30 million Pre-A funding was completed across two investment rounds.
The financing includes:
- $8 million raised in the initial round announced in March
- $22 million secured in the latest funding announcement
The newly raised capital will be used to:
- Expand multimodal data collection across Southeast Asia and North America
- Scale deployment of HOMIE wearable devices
- Advance AI research and engineering
- Grow the U.S.-based engineering team
- Strengthen the company’s data platform capabilities
Investor Confidence in Ropedia’s Vision
One of Ropedia’s angel investors, a research scientist from Amazon, praised the company’s rapid execution and long-term vision.
The investor noted that the founding team combines deep technical expertise with strong execution capabilities, positioning Ropedia to become foundational infrastructure for the emerging physical AI ecosystem.
As robotics companies increasingly require large-scale, real-world datasets, investors believe Ropedia occupies a strategic position similar to how cloud infrastructure companies enabled modern internet computing.
Experienced Founding Team Driving Innovation
Ropedia was founded by three experienced AI researchers:
- Zhaoxi Chen, CEO and Co-founder, recognized for pioneering work in 3D computer vision and multimodal AI.
- Fangzhou Hong, CTO, previously contributed to Meta’s egocentric multimodal intelligence research before advancing work in 3D spatial intelligence.
- Ziwei Liu, Chief Scientist and Associate Professor at Nanyang Technological University (NTU), Singapore.
Founded during the second half of 2025, Ropedia is headquartered in Singapore and also operates an office in Mountain View, California.
Physical AI Could Become the Next Computing Revolution
The rapid emergence of physical AI is creating demand for entirely new categories of infrastructure.
Just as cloud computing depended on data centers and generative AI relied on internet-scale text datasets, robotics companies now require vast quantities of synchronized real-world interaction data to train intelligent machines.
Ropedia aims to become the foundational data infrastructure layer for this next generation of AI by combining wearable hardware, scalable data collection, multimodal processing, and structured datasets into a unified ecosystem.
With its growing customer base, expanding Xperience-10M dataset, and fresh $30 million investment, the company is positioning itself to play a pivotal role in enabling robots capable of understanding and interacting with the physical world.