Mecka AI raises $60M from Sequoia for robot training data

Mecka AI, a startup that collects human motion data to train robots, has raised a $60 million Series B led by Sequoia Capital, TechCrunch reported on October 7, 2026. Nvidia and M12, Microsoft's venture fund, also invested. The round is a bet that robots, like chatbots before them, will need a steady supply of human-made training data.

What Mecka AI does

Mecka pays people to record themselves doing everyday tasks while wearing body sensors, TechCrunch reports. Making coffee and fixing cars are two of the examples. The sensors and a smartphone capture how a person's body moves through each task. Robots can learn physical skills from examples like these, so a large library of recorded motion becomes training material.

AI Weekly reported in June that Mecka records with body sensors and iPhones, and that the company is based in New York. Mecka's own website calls the company "the data and deployment layer for physical AI." Physical AI means AI that controls machines in the real world, such as humanoid robots. The site lists a mobile app for collecting data, web tools for working with it, and a dataset platform called Egoverse.

The round and the earlier money

Mecka was founded in 2024. It raised money twice before this round, both times led by Framework Ventures, according to Crypto Briefing. AI Weekly named Menlo Ventures, SV Angel and Kindred Ventures among the other backers of that earlier money.

Round	Date	Amount	Lead investor
Series A	November 2025	$25 million	Framework Ventures
Extension	June 2026	$35 million	Framework Ventures
Series B	October 2026	$60 million	Sequoia Capital

The Sequoia deal had been reported before it closed. In September, Crypto Briefing and AIChief reported that a Sequoia-led round was taking shape at a valuation of about $500 million, with terms not yet final. TechCrunch's October report cites that figure from its own earlier reporting and does not give a final valuation.

Mecka's website now lists Sequoia, Nvidia, Qualcomm, Microsoft and Samsung as investors.

Why investors want robot data

TechCrunch compares Mecka to Scale AI and Mercor, two companies that supply human-made training data for large language models. Text for chatbots could be gathered from the web. Recordings of people doing physical tasks mostly have to be made on purpose, and that is the business Mecka is in.

The company says the business is growing fast. AI Weekly reported in June that Mecka projected $100 million in annual recurring revenue from contracts it had already signed, with 40 employees. Framework Ventures cofounder Vance Spencer called it "the fastest-growing revenue company that we've ever invested in."

Rivals are raising too. TechCrunch says competitor XDOF was reportedly in talks for a Series B at a $1.2 billion valuation.

What this means for developers

For teams building robots, training data is becoming something to buy rather than collect. Before signing with a supplier like Mecka, ask how the motion is captured and which tasks are covered. Also ask whether the recordings match your robot's body. Motion recorded from a human hand may need work before a two-finger gripper can use it.

Check the rights before you train. Mecka pays people to record themselves, so ask what consent and license terms cover the data. Confirm that you can use it to train models you plan to sell.

Compare suppliers. Sequoia, Nvidia and Microsoft's fund are now behind Mecka, and XDOF is reportedly raising at $1.2 billion. Competition between well-funded suppliers can push data prices down, but it can also tie buyers to one vendor's formats. Ask for a sample dataset from at least two suppliers and test both on your own robot before you commit.

For everyone else, the round shows where robotics money is going. Investors are now funding the supply chain around robots, not only the robots themselves.