Genesis AI— Foundation models and simulation for general-purpose robots.

Genesis AI is a physical-AI lab building a universal robotics foundation model, trained on its own ultra-fast physics simulation engine, to give general-purpose robots dexterous manipulation across everyday tasks.

Visit website
Founded
2024
Headquarters
Paris, France
Legal name
Genesis AI SAS

Highlights

Notable milestones from Genesis AI’s journey — funding rounds, launches, and other moments worth knowing.

launch

Eno rolls out as Genesis's first general-purpose robot

Genesis unveiled Eno, a wheeled, height-adjustable robot with two dexterous arms, built around its GENE robotics foundation model rather than a traditional humanoid body. Targeted production and customer deployments are planned by the end of 2026, first for manufacturing, logistics and laboratories before service and home environments.

Genesis AI's Eno robot placing sample vials into a lab centrifuge.
Eno's first deployment plan starts with industrial customers, including manufacturing, logistics companies and laboratories.
Source
partnership

LG CNS becomes the first enterprise deployment partner for Eno

Genesis and LG CNS set up a strategic partnership to validate Eno in LG facilities before broader industrial rollouts. The deal gives Genesis a real enterprise path for its first robot, starting with environments where LG CNS already runs automation and digital-transformation programs.

Source
launch

GENE-26.5 goes full-stack with a human-like robotic hand

Genesis unveiled its first robotics foundation model alongside its own dexterous, human-sized hand built with partner Wuji Tech — and a matching sensor glove to close the embodiment gap. Demos showed it cracking eggs, playing piano and solving a Rubik's cube mid-air.

A bimanual Genesis robot with two human-like dexterous hands solving a Rubik's cube in mid-air.
GENE-26.5 controlling Genesis's own human-sized hands to solve a Rubik's cube — part of the full-stack demo.
Source
funding

$105M seed co-led by Eclipse and Khosla to build a universal robotics model

Emerging from stealth, Genesis raised one of the largest robotics seed rounds to date, co-led by Eclipse and Khosla Ventures with Bpifrance, HSG, Eric Schmidt and Xavier Niel backing a plan to unify high-fidelity simulation, generative modeling and real-robot data.

Round
Seed
Raised
$105M
Led by
Eclipse, Khosla Ventures
Source
milestone

The simulator goes viral, topping GitHub's trending charts

Within days of release Genesis shot to the top of GitHub trending and lit up Hacker News and AI Twitter, drawing tens of thousands of stars in its first week — it has since passed 29,000 — and turning an academic project into the foundation for a venture-backed company.

Source
launch

Genesis: an open-source physics engine that runs 430,000× faster than real time

The fruit of a two-year collaboration across 20+ labs, Genesis is a pure-Python, photorealistic simulator that runs 10–80× faster than GPU stacks like Isaac Gym and can train a real-world-transferable locomotion policy in 26 seconds on a single RTX 4090 — released fully open source.

A photorealistic Genesis simulation of a bimanual robot stacking a toy on a table in a rendered room.
Genesis pairs a multi-physics engine with a photorealistic renderer to generate synthetic training data at scale.
Source
team

Two Carnegie Mellon roboticists set out to build a robotics foundation model

Zhou Xian, fresh off a CMU robotics PhD, teamed with Théophile Gervet — a former research scientist at Mistral and Skild AI — and Tsun-Hsuan Wang to found Genesis AI across Paris and Palo Alto, betting that simulation, not fleets of real robots, was the fastest path to general-purpose physical intelligence.

Source