Veeda AI, a Canadian startup founded by former Nvidia researcher Sanja Fidler and her team, announced a $90 million seed funding round to advance its multimodal simulated environments designed for training embodied AI agents safely and at scale.
- Veeda AI raised $90M seed round co-led by Khosla Ventures and Radical Ventures
- Focus on multimodal world models to simulate physical environments for AI training
- Seeks to replace real-world robotic trial with scalable simulated reality
What happened
Veeda AI, a newly formed startup headquartered in Canada and led by renowned computer scientist Sanja Fidler, announced today it has raised $90 million in seed funding. This financing round is one of the largest seed investments ever recorded by a Canadian startup and was co-led by major venture capital firms Khosla Ventures and Radical Ventures. Founded just three months ago, Veeda AI is focused on creating advanced multimodal world models that simulate the physical world in immersive, scalable environments.
The company’s technology aims to allow embodied artificial intelligence agents—such as robots—to learn through repeated interactions within these simulated realities rather than relying on the slower, riskier, and less scalable process of trial and error in the real world. This funding injection will accelerate Veeda’s development of simulated environments deemed critical for training next-generation physical AI.
Why it matters
Training robots in the physical environment poses significant challenges: robotic hardware cannot scale the way computational resources can, experiments in the real world carry safety risks, and practical parallelization of experiences is limited. Veeda AI tackles these issues by building world models that enable what Sanja Fidler describes as a “Matrix” for robot learning, providing high-fidelity virtual spaces where robots can safely and efficiently develop their capabilities at scale.
This approach holds great promise for accelerating advances in robotics across industries. High-quality simulation environments could become foundational infrastructure supporting all physical AI applications, from autonomous vehicles to humanoid assistants. The substantial backing from leading venture firms also reflects the growing momentum around simulation and world model technologies as a key frontier in AI research and deployment.
What to watch next
Following Veeda AI’s substantial seed funding, close attention will be on how quickly the company can mature its simulation technology and attract partnerships within the robotics ecosystem. Success could position Veeda as a crucial enabler for physical AI developers worldwide. Monitoring its progress alongside other companies backed by the same investors, such as Waabi Innovation’s autonomous trucking simulations, will provide insight into the evolving landscape of AI-driven simulation technologies.
Additionally, the broader adoption of these simulated realities for training embodied AI may trigger new academic collaborations, regulatory considerations, and competitive dynamics in both Canadian and global AI markets. Industry watchers should also track Veeda’s ability to translate its research during the early Nvidia collaborations into commercial products that deliver measurable benefits for diverse robotics applications.