Canadian AI startup Veeda AI has raised more than $90 million in seed funding to develop world models that can create virtual environments where robots can learn through trial and error.
Veeda AI was founded by former Nvidia AI researchers Sanja Fidler, Zan Gojcic, and Huan Ling. According to The Logic, Radical Ventures and Khosla Ventures back the company, citing people familiar with the matter and corporate filings.
Fidler, who serves as Veeda’s CEO, said robots will ultimately need to learn through interaction rather than relying primarily on imitation. In imitation-based learning, machines are trained to reproduce examples of human or robot behaviour.
“For Physical AI, the central challenge is making this kind of interactive learning possible at scale,” Fidler wrote. “The real world is simply not a practical training ground for robots to learn through trial and error. It is unsafe, expensive, and time-consuming. To scale interactive learning, robots will need to learn in simulated reality.”
Veeda aims to shift a larger share of robot learning into simulated environments. In these virtual settings, robot AI systems can repeatedly perform tasks, make mistakes, and adjust their behaviour without putting physical equipment at risk or requiring large fleets of robots for training.
The company describes world models as the foundational technology behind this approach. These generative models learn from large volumes of sensor and physical-world data to produce simulated environments with the quality, diversity, and physical realism required by robots.
Veeda’s vision is to create a virtual world for Physical AI that resembles a Matrix-like environment, according to the company’s website. Within these simulations, robot AI systems could operate through virtual versions of their bodies and encounter different combinations of objects, tasks, and conditions.
“At Veeda, our sole mission is to build simulated reality for Physical AI,” Fidler added. “Our conviction is that this will become the critical infrastructure layer for all areas of robotics.”
The company’s approach seeks to address one of the major challenges in training physical AI systems: providing robots with enough opportunities to interact with unpredictable environments. By using simulated reality, Veeda aims to enable repeated experimentation at scale while reducing the safety, cost, and time constraints associated with real-world robot training.
The more than $90 million seed funding will support Veeda AI’s efforts to develop its world-model technology and build simulated environments for Physical AI. With backing from Radical Ventures and Khosla Ventures, the startup is positioning its technology as potential infrastructure for the broader robotics industry.
