Genesis AI, a startup developing foundational AI models for robotics, has secured $105 million in seed funding.
The round was co-led by Eclipse Ventures and Khosla Ventures, positioning Genesis as a major player in building general-purpose models for physical automation.
Robotics-focused co-founders
Founded in December by Zhou Xian, a robotics PhD from Carnegie Mellon, and Théophile Gervet, formerly of Mistral AI, the company aims to automate diverse physical tasks using a single AI model — from lab work to domestic chores.
Advertisement
Synthetic data is the key
To sidestep the high costs of collecting real-world robotics data, Genesis AI is leveraging a proprietary physics engine to generate synthetic data. This simulation platform, developed in collaboration with researchers from 18 global institutions, is designed to model the physical world with high fidelity.
Outpacing the competition
Genesis AI claims its simulation tech allows for faster model training compared to rivals relying on Nvidia-based tools. The team now includes 20+ experts in machine learning, robotics, and graphics. Other players in the robotics foundation model space include Physical Intelligence and Skild AI.
Dual hubs in Paris and Silicon Valley
Operating out of Silicon Valley and Paris, Genesis plans to release its robotics model to the broader AI research community by year’s end.
A calculated VC bet
“It’s a big unknown: Will anybody have a large robotics foundation model that will generalize across tasks? That’s a bet we want to go after,” said Khosla Ventures partner Kanu Gulati. “Of all the teams we have seen, we like Genesis’ approach.”
