New York-based Mecka AI, a robotics data startup building the human-motion datasets that physical AI models need to learn dexterous and locomotion skills, disclosed on June 1, 2026 that it has raised $60 million in Series A funding. Framework Ventures led the financing alongside Menlo Ventures, SV Angel, Kindred Ventures and angel investor Ted Xiao, a former Google DeepMind researcher and founding member of Project Prometheus.
Two rounds in one disclosure
The Series A combines two previously unannounced fundraises: a $25 million round closed in November 2025 and a $35 million follow-on closed earlier this year. Mecka founder Charles Gao said the company chose to consolidate the announcement to reflect the company's broader funding posture and signed contracts going into the second half of 2026. Vance Spencer, cofounder of Framework Ventures, called it the fastest-growing revenue company the firm has ever invested in, citing an annualized run-rate target of $100 million.
iPhones, body sensors, and the robotics data gap
Mecka collects hand gestures, walking gaits and full-body kinematics from human teleoperators wearing custom body sensors and iPhones running its capture stack. The data is then converted into the kinds of high-quality demonstration corpora that vision-language-action and world models need to bridge the gap between simulation and the messiness of real-world manipulation tasks. The thesis mirrors what Tesla, NVIDIA, Skild AI and Apptronik have all argued publicly: the bottleneck on physical AI is no longer model architectures but task-relevant data.
Picks-and-shovels for physical AI
The round is part of a broader rotation by venture capital into the picks-and-shovels layer of robotics. Recent rounds in the same neighborhood include Sereact's $110 million Series B, Sunday Robotics' $165 million Series B, and Apptronik's $520 million extension. Investors are betting that data infrastructure companies will capture durable value as the humanoid and industrial robotics arms race accelerates.
Founders and what comes next
Gao and cofounder Mogen Cheng previously sold a payments startup focused on restaurants. Cofounder Jason Chong sold an earlier company to Coinbase, and Duy Nguyen built a sneaker-resale business before joining. The company says proceeds will go to data-collection hardware, hiring across machine learning and operations, and scaling the supply of motion data to robotics customers including humanoid OEMs and foundation-model labs.
Reporting based on coverage from Fortune, Tech Startups, PitchBook and Framework Ventures.