Mecka AI, a fledgling company that gathers and evaluates human motion data to educate humanoid robots and various robotics, is approaching a new funding round led by Sequoia Capital, boasting a valuation of approximately $500 million, as per two sources familiar with the arrangement.

This latest funding comes just three months after Mecka declared it secured $60 million in a round spearheaded by Framework Ventures, featuring contributions from Menlo Ventures, SV Angel, and Kindred Ventures.

TechCrunch has not ascertained the exact amount of the new round. The deal’s specifics are not yet finalized and may still undergo modifications.

Mecka AI has yet to respond to a request for a statement. Sequoia opted not to comment.

Mecka AI was co-founded in 2024 by four entrepreneurs, including Canadians Josh Gao and Mogen Cheng, who earlier established a restaurant fintech business, and Jason Chong, who joined Coinbase following the acquisition of his crypto exchange. Duy Nguyen, the sole non-Canadian member, is responsible for operations at Mecka.

The quartet of co-founders lacks experience in robotics. Nevertheless, they identified a significant gap in physical-world data and acknowledged that capturing real-world interactions was the key impediment affecting general-purpose robots, such as humanoids.

Mecka, which takes its name from “mecha,” a fictional colossal robot commanded by humans, aims to achieve for robotics what Scale AI, Mercor, Surge, and other human data firms have accomplished for LLMs. The startup compensates individuals to document themselves engaging in daily activities — such as brewing coffee or repairing vehicles — utilizing body sensors and smartphones.

As of early June, Mecka anticipated concluding 2026 with an annual run rate of $100 million, Gao informed Fortune at the time the startup announced its earlier fundraising.

Although Mecka AI has not publicly revealed its customer roster, numerous robotics firms and AI laboratories depend on real-world data collected through this “egocentric” methodology, in addition to other physical data gathering techniques like teleoperation, to develop their models.

Other startups amassing real-world data for robot training consist of XDOF, which TechCrunch reported last week was nearing a new round at a $1.2 billion valuation, along with human-data platforms that are expanding beyond LLMs, such as Scale AI and Micro1.