A 25-year-old founder’s solution to the robotics industry's "data famine" has attracted heavyweight investors, signaling a shift from building robots to capturing the data that trains them.
Chinese embodied intelligence startup OriginFlow has raised over 500 million RMB ($69 million) in five months, betting that its novel approach to capturing human movement data can solve a critical bottleneck hindering the deployment of general-purpose robots in homes and factories.
"The industry is facing a universal 'data famine'," according to insights from the company's funding announcement. "The generalization ability of robotic arms has been unable to break through the bottleneck, which is essentially due to the lack of supply of high-quality physical operation data."
The funding, accumulated over Angel, Strategic, and Pre-A1 rounds, was led by Monolith, BlueRun Ventures, and Oasis Capital, with strategic investment from 58.com. OriginFlow's "NeuroScale" technology uses surface electromyography (sEMG) sensors to capture the neural signals behind human muscle movements, a departure from the industry's vision-based "EgoScale" standard that often fails to capture force and tactile feedback.
By providing higher-quality training data, OriginFlow aims to unlock a vast market for robots in non-standardized environments, positioning itself as a key infrastructure provider, or "picks and shovels" play, in the race to build truly intelligent machines. The 500 million RMB investment values this data-centric approach over building the robots themselves.
A New Data Paradigm
The core of OriginFlow's technology, developed by 25-year-old Tsinghua University PhD candidate Qin Shentao, is the "NeuroScale" paradigm. It bypasses the limitations of purely vision-based data collection, which struggles with object occlusion and cannot directly measure the force or tactile feedback crucial for complex manipulation tasks. By tapping directly into nerve signals, the system captures the user's intent and physical interaction with an object.
This approach has the potential to drastically improve robot dexterity, a challenge highlighted by recent developments in the broader logistics automation sector. While companies like Locus Robotics are expanding capabilities through acquisitions to improve grasping, as noted in a recent report, OriginFlow is tackling the problem at the data source. The company claims to have reduced the cost of its sEMG data-collection hardware to the thousand-yuan level (roughly $140), a key prerequisite for mass adoption and large-scale data gathering.
From Lab to Market
OriginFlow is positioning itself not as a competitor to robot manufacturers like Boston Dynamics or Figure AI, but as a crucial enabler for the entire industry. The investor lineup reflects a clear go-to-market strategy. The strategic buy-in from 58.com, a major online marketplace for local services, points to a direct application in domestic settings. OriginFlow can leverage 58.com's network to collect vast amounts of data on high-frequency, non-standard tasks like cleaning, cooking, and sorting, building a valuable skills database for home-service robots.
The significant funding for a data-centric firm mirrors a broader trend in AI, where access to high-quality, proprietary data is becoming a key differentiator. Enterprise AI platform Unframe recently raised $50 million on the back of strong demand from companies looking to move AI projects into production, underscoring the market's appetite for solutions that bridge the gap between AI ambition and real-world execution.
However, as a five-month-old company, OriginFlow faces significant hurdles. The robustness of its technology must be proven in complex, real-world environments like factory floors with electromagnetic interference. Furthermore, it must establish a durable business model that avoids becoming a one-time hardware vendor, ensuring its data and models remain indispensable to large-scale robot manufacturers over the long term.
This article is for informational purposes only and does not constitute investment advice.
