June 26, 2026
Waabi crosses the next frontier of generalization, unlocking truly scalable AVs
Delivering on the promise of Physical AI requires systems that can generalize at human — or even superhuman — levels, transferring learned knowledge to new situations and tasks, performing reliably and safely in the unknown.
There are two distinct frontiers of generalization in self-driving. The first is the ability to generalize across environments and behaviors. This means the virtual driver should be able to apply its knowledge and skills safely across very different Operational Design Domains (ODDs), from highways to dense urban streets, across different traffic patterns, road geometries, and driving behaviours.
This is something the Waabi Driver excels at. As the industry's only verifiable end-to-end AI system with human-like reasoning capabilities, it reached a major milestone in Q1 2025 by expanding beyond highways to safely and confidently operate on complex general surface streets.
The second frontier is the ability to generalize across embodiments. This means the virtual driver should be able to seamlessly transfer to an entirely new vehicle platform with different sensors, control system, and physical characteristics. Historically, this has been one of the most difficult challenges in autonomous driving, with each new platform requiring very extensive engineering, data collection, retraining, and validation.
We built the Waabi Driver from day one as a single AI brain designed to enable true embodiment generalization and automatically adapt to completely different vehicles, just like humans do. When humans get behind the wheel of a new vehicle, we don't have to relearn how to drive. The same should be true for AI.
To prove the Waabi Driver’s ability to generalize in this way, we collaborated with our partners at Volvo Autonomous Solutions to integrate the Waabi Driver onto the Volvo VNL Autonomous. This was a brand new form factor for the Waabi Driver and what unfolded was remarkable.
The Volvo VNL Autonomous powered by the Waabi Driver generalized zero-shot. In other words, without requiring new real-world data, simulation data, or fine-tuning, the Volvo VNL Autonomous, powered by the Waabi Driver, was fully performant from the very first mile. It drove autonomously across both highways and complex surface streets safely, confidently, and smoothly, just like the best human drivers.
With this achievement, the Waabi Driver has set a new benchmark for generalization across the entire industry, proving its ability to transfer learned knowledge across the most complex driving environments and the most challenging form factors. We now know that zero-shot generalization across platforms with different vehicle sizes and shapes, sensor suites and control systems is possible.

Zero-shot transfer of Waabi Driver, trained to drive a Peterbilt 579, and highly performant driving a Volvo VNL Autonomous without requiring new real-world data, simulation data, fine-tuning or engineering. Remarkably the Volvo VNL is an entirely new vehicle platform with different sensors, control system, and physical characteristics.
This marks a new era for autonomous trucking. As the industry moves toward commercialization, the ability to deploy the same AI system across different operating environments and vehicle platforms unlocks the ability to meaningfully scale the technology in a sustainable and economically viable way.
Road testing the Volvo VNL Autonomous, integrated with the Waabi Driver, on public roads is an important proof point of our partnership with Waabi. It also demonstrates the scalability of Volvo’s autonomous truck platform, which is designed to integrate different vehicle models and virtual drivers to enable a wide range of use cases and applications. Together with Waabi, we are advancing autonomous transport solutions toward commercial reality.
Nils Jaeger, President, Volvo Autonomous Solutions
While demonstrated in autonomous trucking, this breakthrough also represents a foundational advance for the broader sphere of Physical AI.
This is a defining moment for physical AI. For the first time in the industry, we have shown that a virtual driver can generalize across fundamentally different embodiments without requiring a single training example —neither real or simulated— or finetuning. This capability has the potential to transform far more than transportation. It is the foundation for a new generation of intelligent machines that can adapt, scale, and operate across the physical world, creating possibilities and opportunities we can scarcely imagine today.
Raquel Urtasun, Founder and CEO, Waabi
The age of truly scalable, adaptable physical AI is no longer a distant vision. It is here, and Waabi is leading the way.