Human Archive, a Silicon Valley-based startup founded by researchers from University of California, Berkeley and Stanford University, announced it has raised $8.2 million in funding to build infrastructure for collecting real-world training data for robots and physical AI systems, according to TechCrunch. The funding round included backing from Wing Venture Capital, NVP Capital, Y Combinator, and angel investors associated with OpenAI, Nvidia, Google, Meta, Mercor, AfterQuery, BAIR, SAIL, and Brad Boa.
The company is partnering with businesses in India’s growing home services, hotel, and restaurant sectors to collect egocentric, first-person video and sensor data for training robots. Workers wear specially designed caps equipped with cameras, along with other devices such as tactile gloves, wrist cameras, and full-body motion-capture suits to capture synchronized RGB-D imagery, motion data, and tactile force data.
Human Archive said it currently has more than 1,000 active headsets deployed across multiple locations and over 50 different devices collecting various forms of training data. The startup believes combining video with motion and tactile information creates more valuable datasets for robotics and physical AI labs than video alone.
The company was founded by Samay Maini, Rushil Agarwal, Shloke Patel, and Raj Patel, with Raj Patel serving as CEO. The founders have backgrounds in robotics, hardware, and tactile data research.
Human Archive said its core thesis is that robotics companies and frontier AI labs face a major shortage of high-quality real-world training data showing humans performing everyday physical work. The company sees India’s rapidly expanding gig economy as a scalable source for collecting that data.
The startup said it works with smaller service providers to offer consumers discounted home services in exchange for agreeing to data collection during visits. Alternatively, customers can pay the standard price for unrecorded service sessions. Human Archive said recordings can also help resolve disputes related to service quality.
The company pays workers a base rate of approximately $1 per hour for participating in data collection efforts. Human Archive said competitors in the space often pay more, but the company believes its local operational presence in India helps lower costs.
Human Archive said its contracts comply with India’s Digital Personal Data Protection Act, adding that workers and consumers are presented with privacy notices and consent disclosures that explain how their data is collected and processed. The company also said all collected footage is anonymized and faces are blurred in recordings.
The startup is also expanding beyond India into Southeast Asia and the United States. Human Archive said it is building a platform that would allow individuals globally to participate in data collection and earn income, while also exploring early pilot programs in the U.S. for services such as cleaning and cooking, tied to robotics data-collection initiatives.
Human Archive said it is additionally developing methods to fine-tune AI models using its internally collected datasets and test those models on robots to demonstrate the effectiveness of the training data to prospective customers.
