Vention has opened a Physical AI laboratory in Montreal focused on developing robotic manipulation technology for manufacturing and moving AI-based robotics research into industrial deployment. The lab is led by Jimmy Li, Vention’s director of Physical AI, and is working on applications for sectors including industrial goods, electronics and automotive manufacturing. Its research covers industrial data collection, robotics control, motion planning, computer vision, vision foundation models, learning from demonstration and reinforcement learning.
Vention said the work is intended to address manufacturing tasks that are complex or unstructured and to test AI-based robotic systems against production requirements including reliability, cost and variability. The company said the lab has already generated intellectual property from its research, without providing further details.
The company plans to use data and feedback from its existing manufacturing customers to support model development and testing. Vention said its technology is deployed at thousands of manufacturers globally, including 90 companies in the Fortune 500.
Physical AI-related revenue at Vention increased 400% over the past year, according to the company. Vention did not disclose the segment’s revenue or its share of total company sales.
The lab currently has a team of eight people and is expanding. Vention said it has more than 16 open positions spanning robotics control, simulation and Physical AI research.
Li, a robotics and machine-learning researcher from McGill University, has more than a decade of peer-reviewed research experience in areas including computer vision, robot perception and autonomous systems. He has led Vention’s Physical AI strategy over the past two years and its technical collaboration with NVIDIA.
One technology developed through the company’s Physical AI work is GRIIP, a modular software pipeline introduced in February 2026. The system covers scene digitalization, object segmentation, pose estimation, grasp selection and collision-free motion planning, using foundation models from NVIDIA alongside Vention’s proprietary models.
Vention plans to release a public GRIIP software development kit that will allow engineering teams to adapt the pipeline to their requirements. The company did not provide a release date.
The company is also working with industrial and electronics manufacturers on robotic applications, including an unnamed large automotive original equipment manufacturer on unstructured tasks in final assembly.
Joelle Pineau, chief AI officer at Cohere, has joined the lab as an external technical adviser. Her role will include advising on model architecture and research priorities and providing connections to the broader AI research community.
Vention founder and CEO Etienne Lacroix said the company is seeking to make Physical AI systems economical and reliable enough for deployment across factories. Li said the lab is structured to incorporate manufacturing customer feedback directly into its research and development process.
