Mecka AI nears $500 million valuation as Sequoia bets on robot training data
The money chasing physical AI just moved from the robots to the data that teaches them. Mecka AI, a two-year-old startup that pays people to record everyday tasks for humanoid-robot training, is reportedly nearing a Sequoia-led round at roughly a $500 million valuation — its second raise in three months.
Mecka AI is closing in on a new round led by Sequoia Capital at a valuation of about $500 million, according to two people with knowledge of the deal — just three months after the startup raised $60 million in a Framework Ventures-led round that included Menlo Ventures, SV Angel and Kindred Ventures. The terms are not final and the round size is undisclosed; Mecka did not respond to a request for comment and Sequoia declined to comment, so this is a reported deal, not an announced one. Even so, the direction is clear: a startup with no announced customers and no robotics pedigree could be worth eight figures more than its June price in a single quarter.
What Mecka actually sells is motion, not robots. The company — founded in 2024 by four co-founders with backgrounds in restaurant fintech and crypto rather than robotics — pays people to record themselves doing ordinary things like making coffee or fixing cars, captured with body sensors and smartphones. The bet is that real-world physical interaction data is the bottleneck holding back general-purpose robots, the same role human-data firms like Scale AI, Mercor and Surge played for language models. As of June, co-founder Josh Gao told Fortune the company projected ending 2026 at a $100 million annual run rate, which would make a $500 million valuation look more like arithmetic than hype.
The pace is the signal. Sequoia is historically patient, and a follow-on round this fast usually means the lead investor saw numbers that justified moving before the market repriced the category. Robot-data startups are having a funding month — Micro1 hit a $500 million run rate, General Intuition raised at a $6 billion valuation, and Generalist reached a reported $3 billion — so Mecka is not an outlier, it is the median of a hot lane. The open question is whether egocentric smartphone-and-sensor capture stays the winning collection method, or whether teleoperation and simulation outcompete it; a $500 million price is a wager on the first answer.
What to watch: the final round size and whether any robotics lab is named as a customer — that single disclosure would tell us whether the run-rate claim holds up outside investor decks.
If you were building a humanoid robot today, would you buy motion data or go collect your own? Tell us in the comments.
