Singapore-based startup Ropedia has raised $30 million in pre-A funding to scale its data infrastructure for physical AI. The company's wearable device, HOMIE, captures multimodal real-world data to train robots and embodied AI systems. The funding will expand data collection and hardware deployment.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
Singapore-based startup Ropedia has announced the completion of a $30 million pre-A funding round. Its core product, the HOMIE wearable device, aims to collect real-world multimodal data to provide training data for physical AI. Whether this model can become the 'data foundation' for the physical AI era remains to be validated by the market.
Ropedia is a Singapore-based startup focused on building data infrastructure for physical AI. The company announced on August 3, 2026, the completion of a $30 million pre-A funding round, conducted in two tranches: $8 million announced via social media on March 16, and $22 million in the latest round. The investors are angel investors with deep experience in AI, robotics, and enterprise technology, as well as long-term financial investors and strategic partners.
The funding will be allocated to three priority areas: expanding data collection in Southeast Asia and North America, scaling HOMIE hardware deployment to support a larger device fleet, and advancing AI research and data platform efforts, including adding engineering roles in the United States.
Ropedia's core hardware product, HOMIE, is a head-mounted wearable device that can simultaneously capture first-person video, audio, depth, hand tracking, gaze, body movement, and camera pose, with timestamps for each data stream. This precise alignment is crucial for physical AI, which requires real-time correlation of perception and action, a capability most existing data collection methods lack.
Ropedia emphasizes that it is a data infrastructure company, not a data labeling service. Standard labeling providers annotate existing data, whereas Ropedia generates its own data, then synchronizes, structures, and continuously optimizes it. Compared to teleoperation-based data collection, HOMIE does not rely on physical robot hardware, so it is not limited to specific robot types. It can be deployed anywhere and worn by anyone, enabling parallel data collection across diverse environments and users.
Ropedia's flagship dataset, Xperience-10M, is claimed to be one of the largest human experience datasets in the industry, containing 10 million interaction episodes and over 10,000 hours of multimodal recordings, encompassing billions of synchronized video, depth, motion capture, and inertial sensor frames. As HOMIE deployment increases, the dataset's coverage of environments, behaviors, and interactions expands, enhancing its value with network growth.
Ropedia operates a 'closed-loop pipeline' that integrates multimodal synchronization and rigorous quality assurance into dataset generation and model alignment fine-tuning. Its business model includes three streams: dataset licensing, selective access to HOMIE hardware, and research collaborations.
Ropedia claims its data collection costs are up to 50 times lower than traditional methods, and HOMIE has entered mass production. The company has served over ten embodied AI and spatial intelligence companies in North America.
An angel investor who is a research scientist at Amazon stated that he invested early in Ropedia because the team possesses deep technical expertise, execution speed, and a clear vision, and he believes Ropedia has the potential to become a foundational company in the physical AI ecosystem.
Ropedia was co-founded in the second half of 2025 by Zhaoxi Chen (CEO), Fangzhou Hong (CTO), and Ziwei Liu (Chief Scientist, Associate Professor at Nanyang Technological University, Singapore). Chen is known for pioneering work in 3D computer vision and multimodal AI, while Hong previously worked on egocentric multimodal intelligence at Meta before contributing to foundational research in 3D spatial intelligence.
Ropedia is headquartered in Singapore with an office in Mountain View, California. The company positions itself as the data infrastructure layer for physical AI, analogous to data centers supporting cloud computing or internet text used to train language models.
This report is primarily based on a press release from Ropedia, which constitutes self-declaration and has not been independently verified. The cost reductions, dataset scale, and other figures mentioned are from the company's official statements, and actual results require further verification.
Ropedia's funding and product showcase a new approach to physical AI data collection: generating real-world multimodal data through wearable devices rather than relying on labeling or teleoperation. However, its claimed cost advantages and dataset value still need market validation, and the competitive landscape for physical AI data infrastructure remains unclear.
Primary report
Primary source