Robot data startup Mecka AI nabs $60M from Sequoia

Mecka AI has secured $60 million in a Series B round led by Sequoia, adding new financial support to a startup working on a less visible component of advanced robotics: data about human movement. Nvidia and M12, Microsoft’s venture fund, joined the financing along with other backers. Mecka AI collects and studies motion information for use in training humanoid machines and other robot designs.

The company’s operating model begins with people rather than robots. It pays individuals to record themselves doing commonplace tasks, then works with the resulting material as motion data. That places the ordinary physical actions of daily life at the center of its business. The company is not presenting the funding as a robotics hardware announcement; its focus is the collection and analysis of information that can help train systems intended to perform physical actions.

The $60 million round is important because it directs substantial new investment toward that layer of the robotics stack. Public attention around humanoid robots often rests on the machines themselves, including their shape, sensors and ability to move. Mecka AI’s stated role is different: it is building a source of examples derived from how people carry out tasks. Those examples may form training material for developers working to give robots a more practical understanding of movement.

Human activity can involve sequences rather than isolated motions. A routine task may require changes of position, coordination, adjustment and response to the surrounding setting. Mecka AI’s effort to obtain recordings of people doing everyday work suggests an attempt to capture that kind of physical context. Its data-collection system also means that contributors are compensated for supplying the raw material, rather than providing it incidentally through a conventional robotics demonstration. The described work is upstream from deployment, not evidence that a particular robot task has already been mastered.

The investment follows an earlier report that the company was nearing a new financing at a $500 million valuation. The completed Series B now confirms that the startup has raised another round, but the information available does not say whether the previously reported valuation was the final figure. A valuation discussed before a deal closes and the eventual terms of that deal can differ, so the earlier number should be viewed as reported context rather than a confirmed outcome of this announcement.

Sequoia led the round, while Nvidia and M12 participated. The investors named in the announcement span established technology and artificial-intelligence interests, and their involvement provides Mecka AI with added capital as it pursues its data-centered strategy. At the same time, the financing alone does not establish how widely its motion datasets will be used, how quickly they will translate into robot capabilities, or which robot makers may rely on the company’s work.

The announcement nevertheless draws attention to a question that can sit behind more visible robotics advances: where relevant training inputs originate. A robot that is meant to operate in human environments may need examples suited to actions performed in those environments. Mecka AI’s answer is paid recording of ordinary activity followed by analysis of the captured movement. The startup says its material is intended for humanoid robots as well as other kinds of machines, broadening the potential scope beyond a single robot category.

Several key details remain undisclosed in the funding news. The available information does not identify Mecka AI’s customers, the amount of data it has gathered, the particular tasks being recorded, or the terms under which the data is supplied and used. It also does not show how the company’s datasets compare with alternatives. Those issues will be central to assessing the effect of the Series B. The next signs of progress will come from whether its human-motion data proves useful in training robots across the applications it says it hopes to serve.

Source: TechCrunch