Mecka AI Raises $60 Million to Train Robots With Human Data
Mecka AI's $60 million Series B, led by Sequoia, backs a bet that human data, not hardware, is what holds general-purpose robots back. Here is how the startup collects it.

Mecka AI, a robotics data startup based in Toronto and New York City, announced on Oct. 7, 2026 a $60 million Series B led by Sequoia Capital. The company sells training data for robots: recordings of people doing physical tasks, turned into material that robot software can learn from. It says the money will go to data infrastructure, its research lab and deploying robots for customers.
Mecka AI closes a $60 million Series B led by Sequoia Capital
According to the company's release, the new investors are NVIDIA, M12 (Microsoft's venture fund), Qualcomm Ventures and Samsung. Existing backers Kindred, Framework Ventures and Neo continue to support it. BetaKit converts the round to about $86 million CAD.
BetaKit adds that several individuals also took part: DoorDash co-founder and CEO Tony Xu, former ServiceNow and Snowflake CEO Frank Slootman, and former Tesla Optimus head Milan Kovac.
Mecka says the capital will scale its data infrastructure, deepen its internal research lab and build out commercial robot deployment. It describes itself as "the data and deployment layer for robotics." Some French headlines cite 55 million euros, which is only a currency conversion of the same $60 million.
Why Mecka says data, not hardware, is the bottleneck for robots
TechCrunch reported on Sept. 11, 2026 that Mecka's founders saw a dearth of physical-world data. They treated capturing real-world interactions as the main bottleneck holding back general-purpose robots, including humanoids.
Training data for a physical robot means examples of how bodies and hands move and handle objects. Chatbots can learn from text already on the internet. Robots have no comparable pile of recordings.
Mecka focuses on egocentric video, meaning first-person footage seen from the worker's own point of view. According to BetaKit, the company bets that this kind of human video helps robots learn faster and scale more efficiently.
How Mecka collects training data from people instead of teleoperation
Mecka pays people to record everyday tasks, such as making coffee, fixing cars and folding laundry, using body sensors and smartphones, according to TechCrunch and Tech Funding News. BetaKit lists the settings: homes, culinary work, chemistry labs, task platforms, metal fabrication and leather shops.
The company designs and manufactures its own capture hardware, Tech Funding News reports. An in-house lab then turns the footage into structured data covering motion tracking, 3D reconstruction and sensor alignment. Per BetaKit, the result is delivered as training-ready data to multiple frontier robotics labs and Big Tech firms.
The approach differs from teleoperation, where a person remote-controls a robot to record demonstrations. Fortune reports that Mecka's core hypothesis is to train robots on data collected from humans rather than through teleoperation. CEO Gao said the idea was mainly explored in research papers before his team partnered with robotics labs to show it could work at scale.
Mecka also installs and runs robots on customer sites
Mecka is also a robot integrator, a company that sets up robots for a client. According to Tech Funding News, it installs hardware, captures data on site, retrains models and runs the system afterward.
This is the "deployment" half of its pitch. The sources give no customer names and no robot counts, so the scale of this business is not known.
What is confirmed and what is Mecka's own claim
Two different $60 million figures circulate. Fortune reported in June 2026 a separate $60 million in Series A funding, made of a $25 million Series A closed in November and a $35 million follow-on led by Framework Ventures. The October announcement is the Series B.
Tech Funding News says total funding is now past $120 million, counting an $8 million seed round listed by Dealroom. Mecka's CEO also posted on X that total funding is over $120M.
Fortune reported in June a projected $100 million annual run rate based on signed contracts, with 40 employees at the time. Tech Funding News notes that this is Mecka's own claim. TechCrunch reported in September that the startup was nearing a round at about a $500 million valuation, and that figure remains unconfirmed. Sources also disagree on when the company began: 2024 according to TechCrunch, 2025 according to Fortune.
| Item | Figure or fact | Source | Status |
|---|---|---|---|
| Series B | $60 million, led by Sequoia Capital | Mecka release, TechCrunch, BetaKit | Confirmed by company |
| Series A (separate) | $60 million total ($25M + $35M) | Fortune, June 2026 | Reported |
| Total funding | Over $120 million | Tech Funding News, CEO post on X | Reported, company-stated |
| Annual run rate | $100 million projected | Fortune, Tech Funding News | Mecka's own claim |
| Valuation | About $500 million | TechCrunch, Sept. 2026 | Unconfirmed |
| Founding year | 2024 or 2025 | TechCrunch, Fortune | Sources disagree |
| Locations | Toronto and New York City | BetaKit, Tech Funding News | Reported |

