With a funding round led by Eclipse and Khosla Ventures, the company—choosing to establish itself in France for its world-class talent—will develop an AI infrastructure focused on general-purpose robotics to accelerate the emergence of generalist robots and the automation of manual tasks.
Genesis AI, a global lab specializing in robotics-focused AI and full-stack robotic products, officially launches today with the mission of enabling the automation of all manual tasks. The company is building a universal foundational model for robotics, along with a horizontal platform dedicated to robotics-focused AI. This announcement comes with a $105 million funding round led by Eclipse and Khosla Ventures, with participation from Alven, Bpifrance, HSG, as well as Eric Schmidt and Xavier Niel.
“We chose to set up in France because of its exceptional pool of talent in artificial intelligence and robotics. The country brings together top-tier engineers, researchers, and developers, which is a decisive advantage to accelerate our development,” said Théophile Gervet, co-founder of Genesis.
While manual tasks represent between $30 to $40 trillion of global GDP, 95% of them remain unautomated due to the limitations of current solutions. Traditional industrial robotics is still dominated by specialized, rigid, and costly systems that are difficult to scale. Genesis aims to revolutionize the next generation of general-purpose robots by providing unprecedented robustness, flexibility, and economic efficiency.
To achieve this, the company is developing a data-centric integrated infrastructure, building a universal data engine that combines high-fidelity physical simulations, multimodal generative modeling, and large-scale real-world data collection. Its simulation platform, entirely developed in-house, will generate high-quality synthetic data, combined with an optimized system for collecting real-world data. This approach enables the training of universal foundational models in robotics, surpassing traditional technical silos.
“While digital AI has made spectacular advances, physical AI—allowing machines to understand and effectively interact with the real world—has lagged behind,” said Zhou Xian, Co-Founder & CEO of Genesis. “By building on the progress made by existing digital AI models, we are bringing human-level intelligence to the physical world. Genesis’s integrated approach paves the way for a new generation of machines capable of automating manual tasks effectively and accessibly. At a time when 75% of global businesses struggle to fill roles, physical AI is more essential than ever.”
Founded by a team of recognized experts from both industry and academia—such as Mistral AI, Nvidia, CMU, MIT, UMD, Stanford, and Columbia (simulation, graphics, robotics, large-scale AI)—Genesis is well-positioned to quickly realize its vision. The company also plans to open-source certain key components of its data engine and foundational model to engage the research community and accelerate progress in the field.
“Today, even in the most automated industrial sectors, the robot-to-human ratio rarely exceeds 1 to 30 due to challenges with dexterity, mobility, and reasoning in real-world environments,” said Charly Mwangi, Partner at Eclipse. “General-purpose robotics represents a major breakthrough with economic potential measured in trillions of dollars. Genesis has the team, the vision, and the strategy to usher in this new era.”
“Physical AI has yet to experience the rise of large language models (LLMs), mainly because collecting real-world data is operationally complex,” said Kanu Gulati of Khosla Ventures. “Genesis’s integrated approach—combining advanced simulations and real-world data in a continuous loop, fully mastering the data pipeline internally—gives them a major strategic advantage. We’re excited to support Genesis from day one.”