Gasgoo Munich-GenRobot.AI has closed its Series A funding round. The round was led by Momenta, a global leader in physical AI, with existing investors doubling down across multiple consecutive rounds.
In just one year since its inception, GenRobot.AI has secured seven consecutive funding rounds, with existing backers doubling down repeatedly. The company has raised over $200 million in total, setting a new record for fundraising in the embodied AI sector—specifically for body-agnostic data. This momentum reflects a shared consensus among strategic partners, capital markets, and the industry on the fundamental logic of "data-driven models."
Image Source: GenRobot.AI
The embodied AI industry is currently at a critical juncture, climbing from "specialized demonstrations" toward "generalized capabilities." Without massive, high-quality data on real human skills, it is impossible to verify whether model routes will converge or to cross the critical threshold of intelligence generalization.
GenRobot.AI was founded to break this bottleneck. As a data-driven physical AI company with a closed evaluation loop, it adheres to the philosophy: "From Human for Human. From Model for Model." The company is dedicated to efficiently converting vast amounts of human skill data into measurable, verifiable capabilities for embodied AI and world models—solidifying the infrastructure foundation for the entire industry.
GenRobot.AI operates on a simple premise: empowering physical AI with human skill data. To that end, it has built an infrastructure that spans human skill data collection, data foundation model training, and large-scale real-machine evaluation and verification.
Earlier this year, Jianzhi released a dataset of over 13,000 hours of body-agnostic data for embodied AI and world models to the global open-source community. That dataset now sees monthly downloads exceeding one million.
GenRobot.AI has developed a proprietary product matrix, including Gen DAS (Data Acquisition System)—a non-intrusive wearable collection device—and Gen Controller for skill terminal data acquisition. By expanding into multiple hardware modalities, the system uses sound, magnetic tactile, piezoresistive, capacitive, flexible fabric, and force feedback sensors to fully record human behavior. This allows for high-precision, high-fidelity collection and reconstruction of body-agnostic data.
The Data Foundation Model, trained in-house by GenRobot.AI, is built specifically for the embodied AI and world model sectors. It covers all dimensions required for model pre-training, including Chain of Thought (CoT), 6D pose, depth information, tactile perception, hand tracking, and full-body mesh. The company is also continuously expanding its standardized human skill system across "Domain-Scenario-Task-Skill."
As artificial intelligence shifts from "predicting the next token" to "completing the next action," GenRobot.AI has a clear mission: continuously expand the scale of data directly available for embodied and world model training, deepen the new infrastructure for physical AI, and establish verifiable, reusable, and quantifiable industry benchmarks for evaluating real-world model capabilities. This ensures that every evolution of embodied AI and world models stands on a foundation of compounding infrastructure.
