Table of Contents
Mecka has raised a $60 million Series B led by Sequoia Capital, adding NVIDIA, Microsoft's M12, Qualcomm Ventures and Samsung as new backers while it scales its robotics data and deployment business.
Co-founder Josh Gao announced the round on X, saying the financing brings Mecka's total funding past $120 million. Returning investors Kindred Ventures, Framework Ventures and Neo also participated, alongside angel investors including DoorDash founder and CEO Tony Xu, former ServiceNow and Snowflake CEO Frank Slootman, and Milan Kovac, the former head of Optimus at Tesla.
The raise comes as robotics companies race to collect the physical-world data needed to train general-purpose robots. Mecka's core argument is that capable models alone are not enough, and that robots also need large volumes of human demonstration data that capture movement, contact, force, geometry, and task execution in real environments.
Mecka Is Building Training Data for Physical AI
Mecka is the data and deployment layer for robotics. That means the company is trying to collect and process the kind of real-world experience that text, image, and video models were able to learn from online, but that robots have historically lacked at scale.
Operating a full-stack system that includes custom multi-sensor capture hardware, global data collection fleets, and an internal research lab focused on turning raw recordings into usable training signals. That work includes motion tracking, 3D reconstruction, sensor alignment, and other processing steps needed to convert human demonstrations into data that can be used by robotics models.
Mecka’s system can produce sub-centimeter hand-pose accuracy on in-the-wild data. Also pointing to EgoVerse, a human-to-robot transfer study conducted with researchers from Georgia Tech, Stanford, UC San Diego, ETH Zürich, MIT and Meta, as evidence that captured human demonstration data can transfer across labs, tasks and robot platforms.
The broader problem is that language models benefited from enormous volumes of recorded human text, while robots must learn from physical interaction with the world. Picking up an object, opening a drawer, or manipulating a tool involves vision, motion, force, timing, and contact dynamics. Those signals are harder to capture than text and harder to replay across different robot bodies.
Revenue Claims and Investor Interest
Mecka says it serves several top frontier robotics labs and multiple Mag 7 companies. The company also passed $100 million in run-rate revenue in June 2026 and is projecting a $300 million run rate by the end of the year.
The round follows a fast funding cycle. In September, Mecka was nearing a Sequoia-led financing at a valuation of about $500 million, three months after announcing a $60 million round led by Framework Ventures.
The investor list reflects the growing strategic value of robotics data. NVIDIA, Qualcomm and Samsung all sit close to the hardware, chip and edge-device markets that could benefit if robots become a larger computing category. Microsoft's M12 also gives the round another link to the broader AI infrastructure market, where software, cloud, models and data pipelines increasingly overlap with physical automation.
