Mecka AI
Category: Robotics / Embodied AI
Data infrastructure and deployment layer for physical AI, collecting egocentric human motion data via body sensors and iPhones to train robotics models. Mecka AI was founded in 2024. The company is led by Josh Gao. Based in New York City, United States. Team size: 11-50. Total funding raised: $68M. Latest round: Series A. Key investors include Framework Ventures, Neo (Suzanne Xie, Ali Partovi), Menlo Ventures, SV Angel, Kindred Ventures, A-Star Capital, Cade, Timeless VC, Offline VC, Ted Xiao (angel).
- Founded
- 2024
- Headquarters
- New York City, United States
- Team size
- 11-50
- Total funding
- $68M
Value proposition
Data infrastructure for Physical AI — Mecka captures real-world egocentric human data (via custom body sensors and iPhones) and transforms it into high-quality labeled datasets for training robotics and embodied AI models, bridging the gap between lab demos and reliable real-world deployment.
Products and solutions
Egoverse (64K+ episodes of egocentric human activity dataset), iOS Mobile App for distributed data capture, Data Tools platform (browse, visualize, inference APIs), Enterprise evaluation and deployment programs, Custom hardware rigs for high-fidelity motion capture, Custom labeled datasets for specific tasks
Unique value
Mecka is the only robotics data platform built on large-scale egocentric human demonstration data (not teleoperation), combining custom hardware rigs, an iOS app for distributed collection, and a data+deployment integration layer that positions them as the 'Scale AI for robotics.'
Target customer
Robotics labs, frontier AI model teams, and enterprises deploying embodied AI systems — anyone needing real-world task-specific human data for robot learning and deployment.
Industries served
Robotics, Embodied AI, Humanoid robots, Physical AI research, Manufacturing, Commercial robotics deployment
Technology advantage
Proprietary body sensor + iPhone-based data collection rigs; Egocentric (first-person) video capture methodology vs. traditional teleoperation; Integrated data-to-deployment layer serving as ecosystem integrator between hardware, model, and commercial partners; EgoVerse open-source dataset collaboration with Georgia Tech, Stanford, Meta Reality Labs, Scale AI; Gamified data collection via 'daily quests' mobile app for global scale
How they differentiate
Unlike competitors who rely on teleoperation (humans manually operating robots) or synthetic data, Mecka collects real human demonstration data at scale via paid gig-economy operators using body sensors and iPhones. They position as the 'data and deployment layer' — not just a data vendor but an integrator that bridges hardware, models, and commercial deployment. Founded by non-roboticist entrepreneurs who brought fresh perspective to the data bottleneck problem.
Main competitors
Scale AI (data labeling for LLMs, expanding into robotics), Physical Intelligence (robot foundation models, software-first approach), Skild AI (robot general intelligence models, $3.5B valuation), MicroAGI (human data collection via home cleaning recordings)
Key partnerships
1x Technologies (early customer/robotics partner), Georgia Tech (EgoVerse research collaboration), Meta Reality Labs Research (EgoVerse consortium), Scale AI (EgoVerse consortium), Stanford University, UC San Diego, ETH Zürich, MIT (EgoVerse academic partners)
Notable customers
1x Technologies, Multiple Fortune 100 companies (undisclosed), Frontier AI model teams (undisclosed)
Major milestones
2024: Company founded (Carbon Based Technology Corp.), 2025-08: Launched from stealth with $8M seed led by Neo, 2025-11: Closed $25M Series A led by Framework Ventures, 2026-05: $35M follow-on funding (total $60M between Series A + follow-on), 2026-06: Public announcement of $60M raise in Fortune, Projects $100M ARR from signed contracts, Launched EgoVerse dataset (64K+ episodes) in collaboration with Georgia Tech, Stanford, Meta, Scale AI, iOS mobile app for distributed data collection, 40 employees and scaling
Growth metrics
Projecting $100M annual run rate from already-signed contracts; 40 employees as of June 2026; 64,382 episodes in EgoVerse dataset
Market positioning
Often described as 'the Scale AI for robotics' — Mecka is a first-mover in large-scale human demonstration data specifically for embodied AI training. The company projects $100M ARR from signed contracts and was characterized by Framework Ventures as 'the fastest-growing revenue company we've ever invested in.'
Geographic focus
Global (US-headquartered with Canadian roots, data collection worldwide)
About Josh Gao
Co-founded Mentum Group Inc. (ghost kitchen payments software, sold 2023); Ivey Business School at Western University graduate; Cansbridge Fellowship 2019; built AI translation app as side project (2023).
Official website: https://www.mecka.ai