Key Takeaways
- • Mecka AI closed $60 million in total funding across two rounds: a $25 million Series A in November and a $35 million follow-on investment, with Framework Ventures leading both rounds alongside Menlo Ventures, SV Angel, Kindred Ventures, and angel investor Ted Xiao.
- • The startup collects human motion data via custom body sensors and iPhones to train embodied AI systems for robotics companies, positioning itself as a first-mover in the data infrastructure layer for physical AI applications.
- • Mecka AI's founding team—CEO Josh Gao and cofounders Mogen Cheng and two others—transitioned from fintech, having previously built and sold a restaurant payments platform in 2023, bringing non-traditional AI backgrounds to the robotics space.
- • The company projects $100 million in annual revenue and is operationalizing real-world data collection through custom hardware deployed at manufacturing facilities, directly addressing a critical bottleneck in robotics model training.
- • Framework Ventures' leadership of the rounds signals institutional validation of embodied AI as a defensible market, with investors including a former Google DeepMind researcher and founding member of Jeff Bezos's Project Prometheus, underscoring the strategic importance of human-motion datasets to robotics development.
Summary
Mecka AI, a New York-based startup, has raised $60 million to build the data infrastructure layer for embodied AI and robotics. The funding comprises a $25 million Series A closed in November and a $35 million follow-on round, both led by Framework Ventures, a venture capital firm previously known for crypto investments. Additional backers include Menlo Ventures, SV Angel, Kindred Ventures, and angel investor Ted Xiao, a former Google DeepMind researcher and founding member of Jeff Bezos's AI startup Project Prometheus. CEO Josh Gao declined to disclose the valuation at which the capital was raised.
The startup's core business model involves collecting granular human motion data—hand gestures, walking gaits, and other physical activities—through dual channels: custom-designed body sensors and standard iPhones. This data is then packaged and sold to robotics companies to train embodied AI models. Gao and cofounders Mogen Cheng and two others represent an unconventional entry point into AI infrastructure; the team previously built and exited a restaurant payments technology company in 2023, bringing fintech rather than traditional deep learning credentials to the robotics problem.
Embodied AI data collection is not entirely novel. Autonomous driving company Wayve has trained models on visual input from vehicle cameras, and startup MicroAGI recently offered free home cleaning services in New York in exchange for recorded footage of cleaners' movements. However, Gao positions Mecka as a first-mover in systematic, large-scale human motion data collection for robotics. The company is operationalizing this vision at scale, with Gao currently overseeing manufacturing facilities in Shenzhen that produce custom data-capture hardware designed by Mecka.
The timing reflects broader market recognition that robotics represents a significant emerging wave in AI commercialization, and that high-quality training data constitutes a critical bottleneck. Mecka projects $100 million in annual revenue, indicating investor confidence in both near-term scaling and the defensibility of its data moat. Framework Ventures' participation—shifting from crypto-focused bets to embodied AI infrastructure—signals institutional repositioning toward hardware-adjacent AI opportunities.
The startup's core business model involves collecting granular human motion data—hand gestures, walking gaits, and other physical activities—through dual channels: custom-designed body sensors and standard iPhones. This data is then packaged and sold to robotics companies to train embodied AI models. Gao and cofounders Mogen Cheng and two others represent an unconventional entry point into AI infrastructure; the team previously built and exited a restaurant payments technology company in 2023, bringing fintech rather than traditional deep learning credentials to the robotics problem.
Embodied AI data collection is not entirely novel. Autonomous driving company Wayve has trained models on visual input from vehicle cameras, and startup MicroAGI recently offered free home cleaning services in New York in exchange for recorded footage of cleaners' movements. However, Gao positions Mecka as a first-mover in systematic, large-scale human motion data collection for robotics. The company is operationalizing this vision at scale, with Gao currently overseeing manufacturing facilities in Shenzhen that produce custom data-capture hardware designed by Mecka.
The timing reflects broader market recognition that robotics represents a significant emerging wave in AI commercialization, and that high-quality training data constitutes a critical bottleneck. Mecka projects $100 million in annual revenue, indicating investor confidence in both near-term scaling and the defensibility of its data moat. Framework Ventures' participation—shifting from crypto-focused bets to embodied AI infrastructure—signals institutional repositioning toward hardware-adjacent AI opportunities.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.