Robot Data Collection Company Mecka AI Completes $500 Million Valuation Led by Sequoia
According to TechCrunch, citing two informed sources, Mecka AI, a startup that collects and analyzes human motion data to train humanoid and other robots, is nearing completion of a new funding round led by Sequoia Capital, with a valuation of approximately $500 million. The specific amount of funding has not been disclosed, and the terms are not yet finalized and may still change. The company did not respond to requests for comment, and Sequoia declined to comment.
This round comes just about three months after the previous one. In June, the company announced a $60 million funding round led by Framework Ventures, with participation from Menlo Ventures, SV Angel, and Kindred Ventures; this was effectively split into a $25 million Series A round in November 2025 and an additional $35 million later. Including early seed funding, the total raised is approximately $68 million. Angel investor Ted Xiao is a former Google DeepMind researcher and one of the founding members of Bezos's AI project. Framework co-founder Vance Spencer has referred to it as the fastest-growing investment in the fund.
The company was founded in 2024 by four individuals: Canadian Josh Gao and Mogen Cheng, who previously worked in food tech; Jason Chong, who joined after the acquisition of a crypto exchange by Coinbase; and Duy Nguyen, who is responsible for operations. The team does not have a formal robotics background, is headquartered in New York, with some team members in Toronto, and has about 40 to 60 employees. The name is derived from "mecha." The product path involves paying users to perform everyday tasks like brewing coffee or fixing cars using motion-sensing devices and smartphones, creating first-person training data through an internal video understanding lab, targeting the human data layer for large models, similar to Scale AI, Mercor, and Surge, rather than relying solely on teleoperation. The client list has not been disclosed, but Gao stated in June to Fortune that contracts have been signed to support an annual run rate of approximately $100 million by the end of 2026.
The robotics lab and cutting-edge model group both require real-world actions. First-person data collection and teleoperation run in parallel, creating a data bottleneck for general-purpose robots. Sequoia's entry at this time signifies a valuation of "body data" on par with language data at private equity prices.
The market mechanism is a scramble for training data. Buyers include labs and funds that need to feed action data to humanoid robots; the sell side is Mecka using the $60 million contract narrative from three months ago to justify a $500 million valuation. Funds flow from Sequoia's accounts into sensors, compensating videographers, and the New York-Shenzhen production line. Beneficiaries are data intermediaries that can sell kitchen and factory actions as subscriptions; those under pressure are collection teams still billing by teleoperation hours and unable to provide first-person scale. The event driver is the leak of undetermined terms, not a delivery announcement.
In three months, the valuation has been marked at $500 million, selling not new robots, but videos of how people reach for a cup.
ABAB AI Insight
After companies like Scale have industrialized language data, body data has become the next layer of scarcity. The four non-robot founders of Mecka have exchanged their experiences in food payment and crypto exits for "who will shoot, how to label, and sell to whom." Framework's transition from crypto ledgers to physical AI indicates a shared risk appetite: models will move from screens into kitchens. Sequoia's involvement at the undetermined terms stage is a brand endorsement for the $500 million valuation, rather than a signed deal.
The capital path prioritizes run rates while keeping client lists confidential. The $100 million guidance is based on undisclosed contracts, with private equity using the term "signed" to elevate the price from tens of millions to $500 million. Funds are directed towards motion-sensing kits, task-based payments for videographers, and labs that convert raw video into strategic data. The Shenzhen factory appears in the founders' itinerary, indicating that the collection hardware is also being mass-produced. If clients are a few leading labs, the valuation is extremely sensitive to single renewals; if it has diversified across multiple clients, Sequoia is buying into the industry's pipeline.
The analogy is that Scale AI provides annotations for large models, and teleoperation farms provide demonstrations for humanoid robots. The difference is that the first-person perspective aims to reduce the unit cost of "humans remotely operating robots," shifting data production from labs to homes and workshops. The industry phase is expanding the data collection network, not yet controlling whole machine shipments. The humanoid craze has turned "lack of data" into a fundable statement; whoever hoards actions first will have bargaining power in the next round of whole machine financing.
The structural change is a reconstruction of the industry chain. The mechanism is that robotic capabilities no longer depend solely on motors and models, but on having a sufficiently diverse distribution of human actions. Teleoperation ties humans to controllers; motion sensing combined with smartphones turns humans into walking sensors. Pricing power shifts from whole machine brands to data wholesalers. Sequoia's entry indicates a belief that this wholesale chain will produce platform-level companies like text annotation.