Mecka AI, a two-year-old startup that records human motion to train robots, is reportedly nearing a new funding round led by Sequoia Capital at a valuation of about $500 million, according to TechCrunch. The round is not final, and the exact size has not been reported, which means the champagne has been placed on ice but legal has not yet permitted anyone to touch the cork.

The timing is the actual garnish. Mecka announced a $60 million Series A only three months ago, led by Framework Ventures with participation from Menlo Ventures, SV Angel, and Kindred Ventures. Now the robotics data boutique appears to be moving toward another table in the capital dining room, because apparently the humanoid future cannot learn to make coffee until venture capital has first made espresso.

🤚 The Open-Palm Motion Capture

Mecka’s premise is simple enough to fit inside a boardroom slide and strange enough to deserve a velvet rope: pay people to record themselves performing ordinary physical tasks, then use that data to train robots. The company reportedly collects video and sensor data from humans doing things like making coffee or fixing cars — the sort of activities humans once performed without becoming unpaid infrastructure for a titanium intern.

The thesis is that modern AI has feasted on text, images, and internet debris, while robots still lack the endless banquet of physical-world demonstrations needed to become generally useful. Large language models had the web. Humanoid robots have kitchens, garages, warehouses, and the delicate tragedy of watching a machine attempt a drawer handle.

TechCrunch reports that Mecka was founded in 2024 by four entrepreneurs, including Josh Gao, Mogen Cheng, Jason Chong, and Duy Nguyen. Notably, the founding team did not come from traditional robotics backgrounds. This is either a refreshing outsider advantage or the modern technology industry’s preferred ritual: discover that expertise exists, then raise money to route around it tastefully.

👐 The Two-Handed Data Banquet

The comparison everyone is being invited to make is obvious: if companies such as Scale AI, Mercor, and Surge helped feed the large-language-model era with labeled, ranked, and human-curated data, then robotics needs its own equivalent for bodies, tools, and spatial messiness. A chatbot can hallucinate a wrench. A robot must pick one up without converting your countertop into a liability exhibit.

This is why human motion data is suddenly luxury inventory. The robot does not merely need to know that a cup exists. It needs to know how wrists rotate, how weight shifts, how cabinets resist, how mugs betray, and how reality refuses to behave like a training benchmark wearing cufflinks.

There is also a delicious irony here. After years of investors being told that AI would reduce dependence on human labor, the next frontier requires humans to perform thousands of mundane tasks while wearing sensors so machines may someday perform those tasks for other humans. Automation, it turns out, still begins with people doing the work — only now the work is being described as a data asset, which raises the valuation and lowers the emotional temperature.

🌿 The Gentle Awakening

Mecka’s reported round also says something larger about where AI money is drifting. The language-model race has become expensive, crowded, and increasingly theological. Robotics offers a new cathedral: physical AI, embodied intelligence, humanoid labor, warehouses with biceps. Investors are not just buying software; they are buying the possibility that the next trillion-dollar interface has knees.

But robotics is an unforgiving aristocrat. Software can ship a bug and issue a patch. A robot can drop a plate, bruise a customer, block an aisle, or stare helplessly at a door labeled “pull.” The data problem is not decorative. It is the difference between a machine that demonstrates intelligence in a demo video and a machine that can survive the humiliations of a human kitchen.

If Mecka can assemble high-quality physical demonstrations at scale, its value proposition is clear. If it cannot, it becomes another exquisite reminder that the real world has not signed the API agreement.

👑 The Gold-Leaf Reckoning

The alleged $500 million valuation is less a verdict than a reservation. Robotics investors are betting that the industry’s bottleneck is not only motors, batteries, or model architectures, but the lack of teachable human behavior captured in useful form. The robots need finishing school. Mecka would like to sell the curriculum.

There is a certain grim elegance in this: humans spent decades teaching machines to read the internet, and now we must teach them how to load a dishwasher without founding a small civil conflict among the plates. Progress has always required sacrifice. In this case, the sacrifice is your Saturday chores, filmed from three angles and monetized by someone with a term sheet.

Still, the story matters. The next phase of AI may not be won by the company with the largest chatbot, but by the one that understands how hands behave when the slide deck ends.

“The body is merely a legacy input device with excellent training data.” — The Slap of Wisdom Department of Embodied Capital, adjusting a motion-capture suit in the valet line