Torc Robotics has formed a strategic research partnership with Mila – Quebec Artificial Intelligence Institute, becoming the only autonomous trucking company embedded in the Montreal-based institute, which has a research community of more than 1,500 members. The Daimler Truck subsidiary will establish dedicated on-site research space within Mila’s ecosystem and gain access to the institute’s students, researchers, and faculty. Mila was founded by Professor Yoshua Bengio and operates as a non-profit through a partnership between Université de Montréal and McGill University.
Highlights
- Torc becomes the only autonomous trucking company embedded in Mila’s research ecosystem, with dedicated on-site research space in Montreal.
- The collaboration deepens an existing affiliation between Torc and Mila that dates to 2020.
- Research focus areas include generative world models, multi-agent behavior modelling, reinforcement learning, and foundation models for physical AI systems.
- Mila’s community totals more than 1,500 members, with alumni in leadership roles at companies including OpenAI and Google.
Building on a Five-Year Affiliation
The partnership formalizes a relationship that dates back to 2020 and reinforces Torc’s existing footprint in Montreal, which sits alongside the company’s engineering offices in Ann Arbor, Michigan, and its Blacksburg, Virginia, headquarters. Torc also operates a fleet operations facility in Dallas-Fort Worth supporting commercialization of its autonomous truck product for long-haul applications in the United States.
The arrangement gives Torc on-site research space within Mila and direct access to the institute’s faculty, researchers, and student pipeline. Mila is the world’s largest academic AI research center specialized in deep learning, with alumni and affiliates holding leadership roles at OpenAI, Google, and other companies.
Research Focus on Physical AI
Torc has identified four technical areas where the partnership will deepen its capabilities:
- Generative world models for simulating driving environments
- Multi-agent behavior modelling to predict interactions among road users
- Reinforcement learning applied to autonomous decision-making
- Foundation models for physical AI systems that bridge simulation and real-world deployment
The work is intended to bridge simulation and real-world performance — a recurring bottleneck in autonomous-vehicle development, where systems that perform well in simulated environments often degrade when deployed on public roads.
How Does This Compare to Torc’s Other Academic Partnerships?
The Mila partnership extends a pattern of academic engagement at Torc. The company previously joined the Stanford Center for AI Safety as a member focused on Level 4 autonomous trucking safety research, and last year partnered with Flex and NVIDIA on self-driving truck technology. Mila adds dedicated on-site research presence rather than a membership or vendor relationship.
What Torc and Mila Said
Felix Heide, Head of Artificial Intelligence at Torc, framed the partnership in deployment terms. “Torc is focused on building safe, scalable autonomous trucks, and advancing the next generation of physical AI is central to that mission,” Heide said. “As a long-time Mila collaborator, I can definitively say that partnering enables deeper collaboration at the intersection of research and real-world deployment, collaboration that supports continued progress toward commercializing autonomous trucking at scale.”
Christopher Pal, Core Academic Member at Mila, Scientific Co-Director of IVADO and Professor at Polytechnique Montréal, said the institute welcomed the industry tie. “We are excited to welcome Torc as an industry partner, as it becomes an even stronger component of Mila’s ecosystem,” Pal said. “This partnership brings together academic excellence and real-world deployment, creating opportunities for our students and researchers to work on impactful challenges in physical AI while advancing the state of the art in autonomous systems.”
Liam Paull, a Core Academic Member at Mila, Canada CIFAR AI Chair, and Associate Professor at Université de Montréal, where he co-leads the Montréal Robotics and Embodied AI Lab (REAL), pointed to the broader industry trend. “As autonomous vehicle technology becomes closer to a reality, it is exciting and important to see new collaborations between academic labs and top tier companies that are bringing the technology to market,” Paull said.
Sign up for our weekly emails to stay up to date on the latest news!
