Mecka AI is closing in on a new funding round led by Sequoia Capital at a valuation of roughly $500 million, according to TechCrunch AI, which cites two people with knowledge of the deal. The startup collects human motion data to train humanoid robots. What makes this move stand out is the timing: the raise comes just three months after Mecka announced a $60 million round led by Framework Ventures, with Menlo Ventures, SV Angel, and Kindred Ventures joining in.
TechCrunch AI reports the exact size of the new round isn’t clear yet, and the terms aren’t final. Mecka didn’t respond to a request for comment, and Sequoia declined to comment.
What Mecka actually does
Mecka pays people to record themselves doing ordinary tasks. Making coffee. Fixing cars. Everyday physical work, captured with body sensors and smartphones. That footage becomes training data for robots that need to understand how humans move and interact with the physical world.
The name comes from “mecha,” the fictional giant robots piloted by humans. The pitch is simple: do for robotics what Scale AI, Mercor, and Surge did for large language models. Those companies built businesses supplying the human-labeled data that made LLMs work. Mecka is betting the same bottleneck exists in robotics, just with physical-world data instead of text.
Here’s what’s interesting about the founding team. Mecka was started in 2024 by four entrepreneurs, and none of them have a robotics background. Canadians Josh Gao and Mogen Cheng previously built a restaurant fintech startup. Jason Chong joined Coinbase after it bought his crypto exchange. Duy Nguyen runs operations. What they spotted was the shortage of real-world data holding general-purpose robots back.
Why this matters
Robots have a data problem that text-based AI already solved. LLMs trained on the open internet, a nearly bottomless pile of text. Robots don’t have that luxury. There’s no giant public library of “how a human picks up a mug” recorded from the right angle with the right sensors.
That gap is exactly what Mecka is filling with its “egocentric” approach, capturing tasks from the human’s point of view. Many robotics companies and AI labs lean on this kind of data, TechCrunch AI notes, alongside other methods like teleoperation where humans remotely puppet the robots.
The money tells the story. A jump from a $60 million round to a reported $500 million valuation in three months signals how hot this corner of AI has become. Mecka projected it would end 2026 at a $100 million annual run rate, Gao told Fortune during the last raise. Investors are clearly buying the thesis that physical-world data is the next scarce resource.
The bigger race
Mecka isn’t alone. The field is filling up fast:
- XDOF, another real-world data startup, was nearing a round at a $1.2 billion valuation, TechCrunch AI reported last week.
- Scale AI and Micro1, both known for LLM data work, are expanding into physical and robotics data.
- Teleoperation shops and sensor-based capture firms are all chasing the same demand from robotics labs.
That’s a lot of capital flowing into one bet: whoever owns the physical-world data pipeline owns leverage over the entire humanoid robot boom.
What to watch next
The deal terms could still change, so treat the $500 million figure as a moving target for now. If it closes, expect the funding race among data-collection startups to heat up further, and expect robotics labs to keep paying a premium for high-quality motion data.
For anyone building or investing in robotics, the signal is clear. Data collection has become its own category, and the companies solving it are getting valued like infrastructure, not vendors. Watch whether Mecka can actually hit that $100 million run rate, because that’s the number that justifies the price.
More details are available at the original TechCrunch AI report.