Cambridge, UK firm incorporated 2022 by ex-Nvidia engineers; builds the Vlab GPU robot simulator and Vlearn framework, backed by a $21.5M seed.
Vsim is a robotics simulation and robot-learning company headquartered in Cambridge, England, registered as VSIM TECHNOLOGY LTD (company number 14509090, incorporated on 28 November 2022) with its office at the Old Swiss building on Cherry Hinton Road. It was co-founded by Michelle Lu and Kier Storey, both former Nvidia engineers who worked on physics simulation technology. Coverage of its 2024 funding placed the company in Manchester; its website and registry record now give Cambridge, where it is hiring on-site engineers.
The company describes itself as a full-stack AI lab for robotics. Its stated goal is on-demand learning: robots that acquire new skills and adapt to unforeseen situations by reasoning with accurate internal physics models and efficient learning algorithms, rather than being fully pre-trained. The product line has three layers. Vlab is a GPU-accelerated robotics simulator with contact and friction models, joint and actuation models, exact CAD mesh simulation, soft objects and soft grippers, cloth, cables, fluids and granular materials, and simulated tactile sensors, IMU, LiDAR and RGB and depth cameras; it supports one-click calibration of simulated models from real-world data and zero-shot transfer of trained policies from simulation to real robots. Vlearn is the robot machine-learning framework built on top, with graphical authoring of robots and environments, import of URDF and MJCF files with USD support announced, visual debugging of inertia, joint limits, collision meshes, contact points and lidar rays, demonstration capture through VR controllers or hand and finger tracking for humanoids, digital-twin testing against a robot's real control commands, and calibration of joint gains, friction and camera placement. Vbot, listed as coming soon, is an on-robot runtime meant to evaluate outcomes of candidate actions in simulation faster than real time and to learn new behaviours at the edge.
Vsim's job openings show that it also trains and deploys learned policies on its own physical robot hardware, covering reinforcement learning, imitation learning, domain randomisation and system identification. Earlier material from the company described its physics engine as a component-based, extensible architecture also aimed at visual effects, digital humans and industrial simulation, but the current site concentrates on robotics.
In September 2024 Vsim announced a 21.5 million US dollar seed round led by EQT Ventures, with Factorial Funds, Samsung Next, Tru Arrow, Xora, IQ Capital, Koro Capital, Concept Ventures, the Lakestar Scout Fund and angel Carles Reina participating. Together with about 2.5 million dollars raised earlier, total funding was reported at 24 million dollars. Michelle Lu discussed the company's approach to accelerating robot learning in episode 164 of a robotics podcast in 2026.
Primary type & automation activities this supplier delivers:
GPU-accelerated multi-physics robotics simulator with sensor simulation and sim-to-real calibration.
Robot machine-learning framework for authoring environments, capturing demonstrations and calibrating digital twins.
On-robot runtime that simulates candidate actions faster than real time and learns new behaviours at the edge.
Contact Vsim
WEBSITE
https://www.vsimrobotics.aiHEADQUARTERS
United Kingdom
Company Facts
Founded
2022
Primary Role
Software/Algorithm
Company Size
-
Primary Region
Europe
Annual Sales
-
Funding Stage
Early-Stage VC
Funding Total
$ 24,000,000
Keywords