Mecka AI
About
About Mecka AI
Mecka designs multi-sensor capture hardware and runs collection operations in homes and commercial environments. Its computer-vision models reconstruct motion, hand poses and geometry, turning human demonstrations into structured training signals. EgoVerse provides first-person interaction data, while its mobile app and data tools support collection, dataset exploration and model evaluation.
Enterprise programmes combine task-specific data capture with model adaptation, robot integration and ongoing operations. Customers can commission a collection programme or work with Mecka on deployment in a particular environment. The company connects robot hardware, learning models and commercial users through this data and integration work.
Articles
Supply
Supply chain relationships
- 1X Technologies
Outbound - Data - Mecka AI supplies human-activity training data for 1X robot-learning programme
Source
Funding
- 2026-10-07 - Series B - $60M - Mecka AI raises $60 million Series B led by Sequoia Capital
The financing supports data infrastructure, internal research and commercial robot deployment. - source: mecka.ai
- 2026-06-01 - Series A and follow-on - Mecka discloses US$60 million across Series A and follow-on financing
Mecka disclosed US$60 million across a US$25 million Series A closed in November and a US$35 million follow-on investment, both led by Framework Ventures. - source: fortune.com
- 2025-08-12 - Seed - $8M - Mecka AI raises $8 million seed round led by Neo
The financing supports collection and processing of human demonstrations for robotics training. - source: upstartsmedia.com
Investors
- Framework Ventures - Lead - VC - 2 funding events - repeat backer - first seen 2026-06-01 - latest 2026-10-07
- Neo - Lead - Unknown - 2 funding events - repeat backer - first seen 2025-08-12 - latest 2026-10-07
- SV Angel - Participant - Angel / Individual - 2 funding events - repeat backer - first seen 2025-08-12 - latest 2026-06-01
- A-Star Capital - Participant - VC - 1 funding event - first seen 2025-08-12
- A* - Participant - Unknown - 1 funding event - first seen 2025-08-12
- Cade - Participant - Unknown - 1 funding event - first seen 2025-08-12
- Cade Ventures - Participant - VC - 1 funding event - first seen 2025-08-12
- Frank Slootman - Participant - Angel / Individual - 1 funding event - first seen 2026-10-07
- Kindred - Participant - Unknown - 1 funding event - first seen 2026-10-07
- Kindred Ventures - Participant - VC - 1 funding event - first seen 2026-06-01
- M12 - Participant - Unknown - 1 funding event - first seen 2026-10-07
- Menlo Ventures - Participant - VC - 1 funding event - first seen 2026-06-01
- Milan Kovac - Participant - Angel / Individual - 1 funding event - first seen 2026-10-07
- NVIDIA - Participant - Unknown - 1 funding event - first seen 2026-10-07
- Offline VC - Participant - Angel / Individual - 1 funding event - first seen 2025-08-12
- Qualcomm Ventures - Participant - Corporate / CVC - 1 funding event - strategic/corporate - first seen 2026-10-07
- Samsung - Participant - Corporate / CVC - 1 funding event - strategic/corporate - first seen 2026-10-07
- Sequoia Capital - Lead - VC - 1 funding event - first seen 2026-10-07
- Ted Xiao - Participant - Angel / Individual - 1 funding event - first seen 2026-06-01
- Timeless - Participant - Unknown - 1 funding event - first seen 2025-08-12
- Tony Xu - Participant - Angel / Individual - 1 funding event - first seen 2026-10-07
Team
- Josh Gao - Co-Founder and Chief Executive Officer - He leads Mecka’s commercial and financing strategy around physical data for robotics training, evaluation and deployment.
- Jason Chong - Co-Founder and Chief Technology Officer - He leads technical development for Mecka’s physical-AI data and robot-training infrastructure.
- Duy Nguyen - Co-Founder, Operations - His role covers the operating model for large-scale data collection and physical-AI training workflows.
- Mogen Cheng - Co-Founder and Chief Product Officer - He leads product work for Mecka’s physical-AI data and deployment platform.
Research
- EgoVerse research release introduces large-scale egocentric human dataset for robot learning
2026-04-08 - Research Publication - EgoVerse introduced a collaborative platform for human data-driven robot learning with 1,362 hours of egocentric human demonstrations. The release covers 80,000 episodes across 1,965 tasks, 240 scenes, and 2,087 demonstrators, with manipulation-relevant annotations. The study reports that policy performance generally improves with more human data, while emphasizing that human data must align with robot learning objectives.
Source
Timeline
Events
- 2026-10-07 - Funding - Mecka AI raises $60 million Series B led by Sequoia Capital
NVIDIA, M12, Qualcomm Ventures and Samsung join returning institutional backers. Tony Xu, Frank Slootman and Milan Kovac also participated.
Source - 2026-09-11 - Market Signal - Mecka AI reportedly nears Sequoia-led financing at $500M valuation
The proposed round remains under negotiation, with the amount not yet reported and terms subject to change. The talks come roughly three months after Mecka disclosed a $60 million financing. Mecka collects real-world human-motion data used to train humanoid and general-purpose robot models.
Source - 2026-06-29 - Acquisition - Mecka AI acquires Docula for robotics data-processing layer
The acquisition adds Docula technology for processing large-scale motion data used in robot training. All three Docula team members joined Mecka AI.
Source - 2026-06-01 - Funding - Mecka discloses US$60 million across Series A and follow-on financing
Framework Ventures led both financings, with participants including Menlo Ventures, SV Angel and Kindred Ventures. The collection platform uses body sensors, phones and custom hardware to capture human motion and task demonstrations for robot learning.
Source - 2026-04-08 - Research Publication - EgoVerse research release introduces large-scale egocentric human dataset for robot learning
EgoVerse introduced a collaborative platform for human data-driven robot learning with 1,362 hours of egocentric human demonstrations. The release covers 80,000 episodes across 1,965 tasks, 240 scenes, and 2,087 demonstrators, with manipulation-relevant annotations. The study reports that policy performance generally improves with more human data, while emphasizing that human data must align with robot learning objectives.
Source - 2025-08-12 - Funding - Mecka AI raises $8 million seed round led by Neo
SV Angel, A-Star Capital, Cade, Timeless and Offline VC participated in the round. Custom collection programmes use dedicated recording rigs; smartphone users can contribute everyday task demonstrations.
Source
Relations
Relationships
- Docula
Acquired Company - Mecka AI acquired Docula to add motion-data processing capability for robot-training workflows.
- 1X Technologies
Partner - Mecka AI has worked with 1X Technologies as an early robotics customer for human activity data used in robot learning.
Context
Comparable profiles
- Noitom
stack adjacency - Embodied Data Capture
- Noraxon
stack adjacency - Embodied Data Capture
- Noitom Robotics
stack adjacency - Embodied Data Capture
- Vision Lab
stack adjacency - Embodied Data Capture · Physical AI Data Infrastructure
- LabyrinthAI
stack adjacency - Physical AI Data Infrastructure
- shotwell.ai
stack adjacency - Robotics Data Infrastructure Layer
- micro1
stack adjacency - Physical AI Data Infrastructure
- Yitai Embodied
stack adjacency - Physical AI Data Infrastructure
- Sureform
stack adjacency - Embodied Data Capture · Physical AI Data Infrastructure · Robotics Data Infrastructure Layer
- Lightwheel
stack adjacency - Physical AI Data Infrastructure · Robotics Data Infrastructure Layer
- microagi
stack adjacency - Embodied Data Capture · Physical AI Data Infrastructure
- Encord
stack adjacency - Physical AI Data Infrastructure · Robotics Data Infrastructure Layer
Shared workflows
- Innodata
workflow adjacency - Human Motion Data Capture
- NC AI
workflow adjacency - Human Motion Data Capture
- TwelveLabs
workflow adjacency - Multimodal Robotics Data Labeling
- Oxford Metrics
workflow adjacency - Human Motion Data Capture
Context tags
- Robotics data infrastructure layer - Stack Layer - Mecka collects and processes robotics training data.
- Physical AI data infrastructure - Stack Layer - Mecka builds data infrastructure for physical-AI robot training.
- Human motion data capture - Workflow - Mecka captures human motion and task data for robotics model training.
- Multimodal robotics data labeling - Workflow - Docula acquisition adds motion-data processing for robot training workflows.
- Embodied data capture - Stack Layer - Mecka’s data collection supports embodied-AI training workflows.