Fundraising Fox

Mecka

9 known investors

Mecka builds a data and integration layer for robotics, providing annotation, context, QA, motion understanding, and evaluation workflows that turn real-world activity into training data for robot learning. It connects hardware, models, and commercial partners, and offers an iOS capture app and a web interface for browsing, visualizing, and querying datasets, serving enterprise robotics teams.

Also known as Mecka Β· Mecka AI

Founders & leadership

JGJosh Gao
Josh GaoinFounder
DNDuy Nguyen
Duy NguyeninFounder

Investors Β· 9

Also in the syndicate Β· 6

A-Star CapitalCadeKindred VenturesNeoleadOffline VCTed Xiao

Company profile

researched Aug 2026

Mecka (Mecka AI) is a physical-AI data company that collects, annotates and delivers real-world human activity data used to train and evaluate robotics and embodied AI models. Rather than relying on teleoperation, in which humans manually pilot a robot, the company's hypothesis is that models can be trained on data recorded from people performing everyday tasks. Data is captured through custom hardware such as body sensors and tactile gloves as well as consumer iPhones, and is then processed with annotation, context labeling, QA, motion understanding and evaluation workflows before delivery to customers. Its large-scale first-person human interaction dataset is branded Egoverse, spanning tasks, spaces and environments and intended for studying how people move, manipulate objects and complete tasks.

The product surface has three parts: an iOS mobile app for collecting environment- and task-specific data in labs and at scale; a web data tool for browsing, visualizing and querying datasets, collaborating with teams and running inference APIs; and an enterprise offering for evaluating physical workflows and supporting real-world deployments. The company states that it does not build robots, positioning itself instead as an integrator that plugs into hardware, model and commercial ecosystems and owns the data and integration layer between them. Reported customer engagements range from custom dataset collection to model evaluations and multi-environment data programs, with the CEO saying the company intends to help firms directly integrate and train robotics models rather than acting only as a data aggregator.

Early data supply combined a consumer-facing component β€” a gig-style model in which users record themselves completing tasks, later reframed as gamified "daily quests" β€” with vetted "higher tier operators" who receive Mecka-built hardware rigs for commissioned corporate training runs. Recorded tasks reported in press coverage include barista work, operating a deep fryer, making a bed and solving a Rubik's Cube.

Founding story

The four cofounders β€” Josh Gao (CEO), Jason Chong (CTO), Mogen Cheng (CPO) and Duy Nguyen (COO) β€” knew each other through mutual friends and came from outside robotics research. Gao, an Ivey Business School graduate at Western University in Ontario, and Cheng previously built and sold a startup providing payments/ghost-kitchen software for restaurants, acquired in 2023; Gao also drew Canadian press in 2023 for an AI translation app built as a side project. Chong was cofounder and CTO of crypto startup Utopia Labs through its acquisition by Coinbase, and Nguyen made money reselling sneakers. After deciding to pursue AI for robots, the team spent months reading research papers, visiting robotics labs and messaging roboticists, concluding that data was the sector's bottleneck. They landed on training robots on human-collected rather than teleoperation data, partnered with robotics labs to demonstrate it at scale, and in 2025 β€” after amassing thousands of gigabytes of data β€” formally started the company. The name derives from "mecha," the piloted robots of the Gundam anime series, with one letter changed.

Business model

Mecka sells real-world human activity data and related services to robotics and AI organizations. Engagements include access to its existing datasets, commissioned task-specific data collection, model evaluation runs, and larger multi-environment data programs, alongside enterprise support for evaluating physical workflows and deploying robots. Supply is sourced from paid human contributors β€” including tradespeople such as baristas, cooks and mechanics β€” who record themselves performing tasks using smartphones or company-supplied hardware, with delivered video anonymized and labeled for model ingestion.

Revenue comes from paid contracts with corporate and research customers for datasets, custom collection, evaluation and deployment services. The company reported an projected annual run rate of $100 million based on already-signed contracts as of June 2026, and stated it holds paid contracts with several Fortune 100 corporations.

Traction

Reported traction includes an early partnership with 1X Technologies, whose World Model generation used household motion data from Mecka; paid contracts with several unnamed Fortune 100 corporations; a projected $100 million annual run rate from signed contracts; and a few thousand app downloads with continuing active users as of August 2025. An investor described the company as the fastest-growing revenue company in its portfolio. Headcount reached 40 employees by June 2026, with 25 open roles listed across six teams.

Latest developments

In June 2026, Fortune reported that Mecka had raised $60 million in total across two previously unannounced fundraises β€” a $25 million Series A closed in November 2025 and a $35 million follow-on β€” led by Framework Ventures, with Menlo Ventures, SV Angel, Kindred Ventures and angel Ted Xiao participating. At the same time the company reported 40 employees, a projected $100 million annual run rate from signed contracts, a New York City headquarters, and custom capture equipment in production at a Shenzhen factory. Its careers page lists 25 open roles across research, platform engineering, hardware and robotics, field engineering, operations and corporate functions.

β–ΈFull profile β€” market position, technology, go-to-market, geography, history, risks & controversies

Market position

Mecka is positioned as a picks-and-shovels data supplier for physical AI, compared in press coverage to the role Scale AI played for large language models, alongside other data-labeling firms such as Invisible, Mercor and micro1. Its CEO claims first-mover status in human-sourced embodied data; coverage notes the category is becoming crowded and that some investors and founders question the cost of operations, the specificity of robotics data needs, and whether AI robotics labs will outsource collection rather than gather data in-house. Other companies cited as working with embodied or real-world captured data include Wayve and MicroAGI.

The company's central bet is that human-collected data β€” first-person recordings of people performing real tasks β€” can substitute for costly robot teleoperation data. It claims to own both sides of the pipeline: the recording (via its own hardware rigs, tactile gloves, body sensors and iOS app) and the downstream analysis, labeling and evaluation. It also frames itself as an integrator rather than a robot maker, offering deployment and integration support in addition to data, and maintains a proprietary large-scale egocentric dataset (Egoverse).

Technology

The stack combines capture hardware and software with a data-processing platform. Capture uses custom-designed equipment β€” body sensors and a tactile gloves program, manufactured via a Shenzhen factory β€” as well as iPhones. Captured video and sensor streams pass through annotation, context enrichment, QA, motion understanding and evaluation pipelines that the company describes as necessary because raw video alone is insufficient for training. A set of "signal and action primitives" or perception modules defines the dataset's signal vocabulary. Outputs are exposed through the Egoverse dataset, a web interface for browsing, visualizing and querying data, and inference APIs. Job listings indicate work in spatial AI and neural reconstruction, RL and simulation, video understanding and world models, hand tracking and manipulation, computer vision and robot navigation, and embedded/edge platforms.

Go-to-market

Direct enterprise and research sales, with website entry points for dataset access, commissioned collection, model evaluations and enterprise deployment programs. Distribution of the collection layer runs through a public iOS app and a web data platform, plus partnerships with robotics labs used early on to validate the human-data approach. Press launch coverage and founder-led outreach to roboticists supported initial awareness.

Robotics labs, frontier AI model teams, and enterprises deploying embodied systems that need task-specific real-world data, evaluation, or support moving from research to deployment. Reported customers and partners include home robotics company 1X Technologies and unnamed Fortune 100 corporations.

Geography

Described at launch in August 2025 as Toronto-based with a mostly Canadian engineering team and operators concentrated in Canada; by June 2026 reported as headquartered in New York City. Hiring spans the United States, Canada and China, with roles in New York, Toronto, Shenzhen and North America generally, and manufacturing of custom capture equipment in Shenzhen.

History

Mecka's earliest data-collection effort was structured as a crypto-inspired, decentralized points program launched in 2024, which some users speculated might become a token; no token was issued, and the company later said the crypto association created friction when selling to larger corporations. The approach shifted to gamified daily quests in a smartphone app, with a few thousand downloads and continuing active users as of August 2025, plus proprietary hardware rigs for trusted operators, initially mostly in Canada. The company launched publicly from stealth in August 2025 as a Toronto-based startup alongside an $8 million seed round. By June 2026 it was described as headquartered in New York City with 40 employees, having closed a $25 million Series A in November 2025 and a $35 million follow-on, for $60 million disclosed in total.

Risks & controversies

Coverage notes skepticism from AI investors and founders about the model's cost, the crowding of the robotics data category, and whether robotics labs will outsource data collection given how specific and fast-changing their data needs are. The founders have no robotics background. The company's early crypto-inspired points program, which some users treated as a prospective token, produced no token and, per the CEO, left residual stigma with larger corporate customers. The business also raises privacy considerations, as it depends on video recordings of people performing tasks in homes and workplaces, which the company says are anonymized before delivery. Valuations for the Series A and follow-on were not disclosed, and customers other than 1X are unnamed.

Compiled by commissioned research from 8 cited public sources β€” announcements, filings, and press listed under research sources below.

Key figures

latest reported
EmployeesJun 202640 people
Hiring teamsJan 20266 teams
Open rolesJan 202625 positions
Projected annual run rate (based on signed contracts)Jun 2026$100M
Total funding raisedJun 2026$60M

Company-reported or press-reported figures, each dated to when it was claimed β€” not independently audited.

Timeline Β· 7

launches, deals, and filings
Jun 2026
Mecka AI discloses $60M raised, including $35M follow-on

Fortune reported that Mecka AI had raised $60 million in total across a $25 million Series A and a $35 million follow-on investment, both led by Framework Ventures with Menlo Ventures, SV Angel, Kindred Ventures and angel Ted Xiao participating.

$60M source β†—

Jan 2026
Hiring across New York, Toronto and Shenzhen

Careers page lists 25 open full-time roles across six teams in the United States, Canada and China, including research, platform engineering, hardware and robotics, field engineering and operations.

source β†—

Jan 2026
iOS data capture app and web data tools available

Mecka offers a mobile app on iOS for capturing environment- and task-specific data, plus a web interface for browsing, visualizing and querying datasets, running inference APIs and viewing the Egoverse dataset.

source β†—

Nov 2025
$25M Series A closed

Mecka AI closed a $25 million Series A round led by Framework Ventures in November 2025; the round was not announced until June 2026 and the valuation was not disclosed.

$25M source β†—

Aug 2025
Mecka AI launches from stealth with $8M seed round

Toronto-based Mecka AI came out of stealth with an $8 million seed round led by Neo, with participation from SV Angel, A-Star Capital, Cade, Timeless and Offline VC.

$8M source β†—

Aug 2025
1X Technologies named as early partner

Home robotics company 1X Technologies was described as an early partner; household motion data from Mecka was used in the generation of 1X's World Model.

source β†—

Jan 2024
Crypto-inspired points program for decentralized data collection

The company launched a points program intended to crowdsource global video data collection; some users speculated it would become a token, but no token was issued and the effort was later replaced by gamified in-app daily quests.

source β†—

Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.

In the news

β–ΈResearch sources Β· 8

primary sources listed

8 public sources were cited for this profile; the first-party ones are listed here.

Frequently asked questions

What does Mecka do?
Mecka builds data, evaluation and deployment infrastructure that trains robots on human activity captured via sensors and iPhones.
Who founded Mecka?
Mecka was founded by Josh Gao, Duy Nguyen.
Who are Mecka's investors?
Mecka's investors include Adverb Ventures, Cade Ventures, Forerunner Ventures.