Mecka raises $60 million led by Sequoia to sell human motion data for robots

Mecka, a start-up based in Toronto and New York that collects data on how people move so that humanoid robots can learn from it, announced on Wednesday, October 7, 2026 that it has raised $60 million in a Series B round led by Sequoia Capital. New backers in the round include NVIDIA, Microsoft's venture fund M12, Qualcomm Ventures and Samsung. Earlier investors Kindred, Framework Ventures and Neo also took part. The company says it will use the money to make more data-capture hardware, expand its research and increase the number of robots it deploys.

Humanoid robots have a data problem. Large language models learned from the enormous amount of text already on the internet, but there is no equivalent library showing a robot exactly how a human hand grips a cup, turns a spanner or folds a shirt. Robot makers either collect that data slowly themselves or buy it. Mecka is betting that a business selling this raw material can grow faster than the robot makers it supplies.

The investor list shows who thinks that bet is sound. NVIDIA, Qualcomm and Samsung all make chips or devices that would sit inside future robots, so better training data supports their own markets. Sequoia leading the round adds a top-tier venture name to a category that many investors still treat as early and speculative. The presence of a former head of Tesla's humanoid programme among the angel investors signals interest from people who have built robots themselves.

The revenue claim is the part to watch. A run rate above $100 million is large for a company at the Series B stage, and the target of tripling it within roughly six months is ambitious. These are the company's own figures, and some analysts have questioned whether the year-end goal is reachable. Readers should treat them as targets, not results.

There is also a labour angle. Mecka's model turns everyday physical skill into paid data work. That could create a new kind of flexible job, but it also raises questions about pay, consent and how the recordings are used once robots are trained on them.

Short-term: Mecka can build more sensor kits, sign up more people to record tasks and supply larger datasets to robot companies. Competing data suppliers will face pressure to raise money or find a niche.

Long-term: If reliable human motion data becomes a traded commodity, robot makers could train machines faster and more cheaply, bringing general-purpose humanoids closer to factories, warehouses and homes. If demand falls short, the company's high revenue targets will be tested.

Who is affected: Humanoid robot developers, chip and device makers, gig workers who could be paid to record tasks, and investors weighing how much to commit to robotics infrastructure. Indian robotics and data-labelling firms, which already supply AI training data to global clients, may see a new line of demand for physical-task data.

Questions and Answers