Performances
RD Popularity Index
RD Growth Index
Comparison
Fundraising
Follow Updates
Follow Lists
People
About Mecka.AI
Mecka.AI produces robot training data: it organizes users to record everyday manipulation and movement on their phones, then cleans, labels and structures the footage into datasets sold to robotics companies, with an early points-and-tasks system that carried a strong crypto flavor.
It targets the scarcity of physical-world data, where language and images are abundantly scraped but the grasping, walking and manipulation footage robots need is hard to collect at scale and teleoperation is expensive. Crowdsourcing instead of demonstrations is a sound idea, with data quality and consistency the hard part. The company later played down the onchain narrative and repositioned as a conventional data provider, and early points and token expectations faded.
Mecka.AI has evolved from a data collection platform into the infrastructure layer for physical AI. Key milestones include the 2026 $60M Series A led by Framework Ventures, which validated its data flywheel model. The acquisition of Vancouver-based Docula marked a strategic shift from collection to high-throughput processing, solving the data normalization challenge at scale. The launch of Egoverse, a large-scale human interaction dataset built from first-person activity across real-world tasks, established its proprietary data moat. Early partnerships with robotics firms like 1x validated the approach of training on human demonstration data rather than teleoperation. The company's pivot to a dual business model—selling training data while building deployment infrastructure—positioned it as an integrator across hardware, model, and commercial ecosystems. Its expansion into the US market with a New York office, while maintaining Canadian roots, reflects its ambition to lead the robotics data layer globally.
Mecka launched a new iOS app on the Apple App Store in February 2026 for first-person data recording, upload and task management. In June 2026 it closed a $35 million Series A extension led by Framework Ventures, bringing total funding to roughly $68 million and pushing the company's valuation toward $500 million; that same month it acquired Docula, a video-understanding lab that processes raw footage into training-ready formats. Disclosed annual recurring revenue run rate has reached about $100 million.
Mecka.AI's forward strategy centers on scaling Physical AI training-data infrastructure: built around its flagship Egoverse dataset, it captures real, unscripted human-motion data with custom sensors and iPhones for robotics labs, frontier model teams, and enterprise customers. The company plans to keep growing revenue through enterprise contracts and to shift from early open-ended reward incentives toward more specialized paid data collection, emphasizing commercial deployment and revenue scale over token mechanics.
None of the four co-founders came from the robotics industry. CEO Josh Gao previously co-founded a restaurant software company with Mogen Cheng and sold the business in 2023; Cheng is now CPO, owning the productisation of data collection and the annotation pipeline.
CTO Jason Chong is a familiar name in crypto, having founded the payments startup Utopia Labs before it was acquired by Coinbase, and he rebuilt Mecka's collection pipeline and quality control after joining. COO Duy Nguyen, who made his early money reselling sneakers, runs the global collection network and operations.