Mecka raises $60M Series B led by Sequoia to scale physical AI data and deployment infrastructure

Date: 2026-10-07
Type: ARTICLE
Topics: Funding, Egoverse, Physical Ai, Robot Learning, Sequoia Capital
Deck: Sequoia leads the financing as Mecka expands robot-learning data, physical evaluation and commercial deployment infrastructure.
URL: https://korthosrobotics.com/research/mecka-series-b-sequoia-physical-ai-data-deployment
Referenced profiles: Mecka AI (https://korthosrobotics.com/ecosystem/company/mecka-ai)

Mecka has raised a $60 million Series B led by Sequoia Capital, bringing disclosed financing to roughly $120 million. The company provides data, evaluation and deployment infrastructure for physical AI, connecting robot hardware, foundation models and commercial users.
The round includes NVIDIA, M12 — Microsoft’s Venture Fund, Qualcomm Ventures and Samsung, with continued backing from Kindred, Framework Ventures and Neo. Angel investors include DoorDash CEO Tony Xu, former ServiceNow and Snowflake CEO Frank Slootman, and former Tesla Optimus head Milan Kovac. Mecka says the capital will expand data infrastructure, its internal research lab and commercial robot deployment.
In June, Mecka disclosed $60 million across earlier financings led by Framework Ventures: a $25 million Series A closed in November 2025 and a subsequent $35 million investment. Menlo Ventures, SV Angel and Kindred Ventures also participated. CEO Josh Gao then projected a $100 million annual revenue run rate by the end of 2026 based on signed contracts, according to Fortune.
Mecka’s latest announcement says it surpassed a $100 million revenue run rate in June and projects $300 million by year end. These are company-reported annualised figures and a forecast, rather than disclosed full-year revenue.
TechCrunch reported September talks at a roughly $500 million valuation. Mecka’s completed-round announcement does not disclose a valuation.
Mecka’s platform combines human activity capture, annotation and motion understanding with model evaluation and enterprise deployment support. It describes itself as a robotics integrator, supplying the data and integration layer between hardware, models and commercial partners. Its deployment work includes on-site data capture, post-training and ongoing operations.
One part of the stack is Egoverse, a collection of first-person human interaction data for robot learning. Its public explorer currently lists approximately 450,000 episodes, more than 4,200 hours of activity and over 28,000 represented tasks, spanning kitchens, workshops, laboratories and homes.
In September, Mecka published WetLabs, a benchmark testing frontier AI models controlling dual-arm robotic systems across nine laboratory tasks. It recorded 540 attempts across GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5, with human evaluators confirming 110 successful runs. The tests used no demonstrations or fine-tuning, and covered labware manipulation rather than liquid handling, sterility or experimental design.