Funds Amassed for Constructing All-encompassing Robotics Template Through Genesis AI, Totaling $105 Million
Genesis AI, a pioneering physical AI research lab and full-stack robotics company, has recently stepped out of the shadows with a bold mission to revolutionize the field of robotics and physical AI. The company, co-led by Eclipse and Khosla Ventures, has secured $105 million in funding, with participation from notable investors including Bpifrance, HSG, Eric Schmidt, and Xavier Niel.
The Genesis AI team, led by CEO Zhou Xian (Ph.D. in robotics, Carnegie Mellon) and Théophile Gervet (former research scientist at Mistral AI and Skild AI), consists of over 20 researchers specializing in robotics, machine learning, and graphics. This expertise stems from a collaborative academic project involving researchers from 18 universities.
Genesis AI's approach is data-centric, focusing on developing a scalable, universal data engine that unifies high-fidelity physics simulation, multimodal generative modeling, and large-scale real robot data collection. This strategy aims to overcome the limitations of existing robotic systems, which are often brittle, inflexible, and narrowly focused.
The company's unique approach involves the integration of synthetic and real data. Genesis AI's proprietary simulation stack generates rich synthetic data at scale, while a robust real-world data collection system provides diversity and quality. This dual-engine strategy bridges the gap between simulation and reality, enabling the training of more robust and general-purpose robotic models.
Genesis AI's vision is to democratize physical labor automation, creating robots that are robust, flexible, and cost-efficient. These robots would be capable of automating a wide range of repetitive tasks, from laboratory work to household chores. The company's universal robotics foundation model (RFM) is intended to be the backbone for generalist robots, enabling them to perform diverse tasks in various environments, much like how large language models have transformed digital AI.
With estimates suggesting that physical labor contributes $30–40 trillion to global GDP and the majority remains unautomated, Genesis AI aims to unlock unlimited physical labor by making automation accessible and scalable across industries. The company's strategy of owning the entire data pipeline in-house gives it a unique data advantage, according to Eclipse Partner Charly Mwangi.
Genesis AI's in-house developed simulation stack will produce rich synthetic data at scale, along with a more efficient and scalable real-world data collection system. This approach could potentially provide the company with a significant advantage in the development of a universal foundation model for robotics.
The company plans to open-source components of its data engine and foundation model, further contributing to the advancement of the field. Genesis AI's vision is to bring human-level intelligence into the physical world by leveraging advances in digital AI and translating them into breakthroughs in physical AI, ultimately automating all physical labor and defining the next major chapter in human productivity.
Artificial Intelligence plays a significant role in Genesis AI's vision, as they aim to develop a universal data engine that incorporates multimodal generative modeling, a key aspect of AI. Furthermore, the company is working on a universal robotics foundation model (RFM), which resembles large language models in digital AI, demonstrating the integration of AI principles in physical AI research.
