Written by Admin Alex · Fact-Checked by M.Ali · Info Verified September 2026
We review and update this article regularly as new information becomes available.
TL;DR: Mecka AI pays regular people to record themselves doing chores like making coffee or folding laundry, then sells that motion data to robotics companies. The startup is reportedly closing a funding round led by Sequoia Capital that would value it near $500 million, just three months after a $60 million round put it on a $100 million revenue run rate.
The Job Is Filming Yourself Doing Chores
Imagine getting paid to strap on a set of sensors, grab your phone, and fold a load of laundry. That is a real job now. A startup called Mecka AI pays ordinary people, not actors, not engineers, just regular people, to record themselves making coffee, fixing a leaky faucet, folding towels, and doing the hundred other small physical tasks nobody thinks twice about.
The footage is not headed to YouTube. It is training data. Robotics companies and AI labs building humanoid robots need to see, in painstaking detail, how a human hand grips a mug or how a wrist rotates to twist a cap. Mecka captures that from the first-person point of view, what researchers call “egocentric” data, using body sensors paired with smartphones. Feed enough of that footage into a model and, in theory, a robot starts learning to move through the physical world the way a person does.
Why This Counts As Hard Tech
It sounds almost too simple to be a venture-backed business. But the bet behind Mecka is that the hard part of building general-purpose robots was never really the software. Language models got good fast because the internet is full of text. Robots have no equivalent library of “how humans move.” Someone has to go collect it, one folded shirt at a time.
That scarcity is now worth real money. Mecka is reportedly closing a new funding round led by Sequoia Capital that would value the company at roughly $500 million. It is a striking number for a two-year-old company whose product, at its core, is paying strangers to film chores.
Four Founders, None of Them Robotics People
Here is the part that makes the story more interesting than another AI funding headline. Mecka was founded in 2024 by four people, and not one of them came from a robotics background. Josh Gao and Mogen Cheng, both Canadian, started the company alongside Jason Chong, who previously worked at Coinbase following its acquisition of a crypto exchange he had built, and Duy Nguyen, who runs operations.
That is an unusual bench for a company now supplying data to humanoid robot makers. It may also explain how fast Mecka moved. The founders were not trying to solve robotics. They were trying to solve a recruiting and marketplace problem: how to find, verify, and pay large numbers of people to generate a specific, valuable kind of data at scale. That is closer to a gig-economy playbook than an engineering one, and it happens to be exactly what this moment in robotics needs.
The Money Moved Fast
Mecka raised about $60 million in June 2026, led by Framework Ventures, with Menlo Ventures, SV Angel, and Kindred Ventures joining in. At the time, the company was already running at roughly a $100 million annualized revenue rate. Three months later, it is reportedly closing a round at nearly five times its prior valuation. Few startups in any category get repriced that quickly. Fewer still do it by asking people to film themselves doing dishes.
Robots Get Their Own Scale AI Moment
None of this is happening in a vacuum. During the large language model boom, a wave of companies including Scale AI, Mercor, and Surge built businesses around labeling and generating the data that trained chatbots. Mecka is essentially that same playbook, pointed at the physical world instead of text.
It has company. XDOF, another robotics-data startup, is reportedly valued around $1.2 billion. Scale AI and Micro1, both veterans of the language-model data business, are pushing into robotics data too. Investors clearly think humanoid robots are close enough to matter, and that whoever controls the training data controls the choke point, the way chip makers did for the last AI wave.
Bottom Line
The strangest thing about Mecka AI is also the most telling thing about where AI is headed next. For a decade, training data meant scraping the internet. Now it means paying a stranger to fold their laundry on camera. That a company built by four people with zero robotics experience could ride that shift to a $500 million valuation in two years says less about them and more about how badly the robotics industry needs raw human movement. Whoever keeps supplying that footage cheaply and at scale, not whoever writes the cleverest algorithm, may end up owning the most valuable real estate in the humanoid robot business.
