Freight doesn't move on a fixed map. Every day, trucks move billions of dollars of goods across a sprawling, constantly shifting network of highways, surface streets, distribution centers, and customer facilities. Lanes open and close with demand, seasons, and contracts. For shippers and carriers, operating flexibly within that network is critical to growth and to being a reliable partner to their customers.
To be useful, autonomous trucks need to go where customers need them, not the other way around. Customers don't need a truck that can drive one route perfectly; they need one that can drive any of their routes safely and reliably. Approaches to date fall short on exactly this point: they do not generalize.
Geographic expansion is painstaking in this industry. Every new lane is a massive effort, worked one at a time through the same painful cycle: collect data on the route, engineer the system to handle everything that route throws at it, test it on the road, find what's still broken, and go back to collecting data again. Round and round for months or even years, before performance is good enough on that single lane. Critically, this process has to happen again for the next lane, and the one after that. The result is a business that scales at the speed of its data collection and re-engineering cycles: slowly and at a significant cost.
This is the core bottleneck standing between where autonomous trucking is today and the freight network the industry actually needs to serve.
Last week, Waabi reached a milestone that unlocks a fundamentally different way to scale — one that changes the economics of the entire autonomous trucking business.
For the first time, the Waabi Driver drove the Dallas–San Antonio lane — one of Texas's most important transportation arteries — with zero new real-world data, zero simulated data, and zero fine-tuning or engineering to prepare the system for the route.
The Waabi Driver completing an autonomous run on the Dallas-San Antonio corridor with zero new real-world data, zero simulated data, and zero fine-tuning or engineering to prepare the system for the route.
This doesn't just expand Waabi's footprint in Texas, where we're rapidly growing commercial operations, it demonstrates a capability critical to scaling self-driving: navigating a completely new route zero-shot. What normally takes months or years to unlock, the Waabi Driver did near-instantly, handling complex driving scenarios in a totally new geography it had never seen before, safely, smoothly, and fully autonomously with zero disengagements.
Unlocking new operating domains instantly has been our vision from day one, and it's exactly what we've been building toward with our breakthrough physical AI platform. Today we prove that this vision is a reality.
This isn't the only axis on which we've demonstrated generalization capability. Earlier, we achieved a zero-shot milestone with Volvo, deploying the Waabi Driver on a new vehicle platform, with a different shape, different sensor placements, and different dynamics than the trucks we'd built and tuned on before. There was no embodiment-specific retraining cycle needed to reach parity from day one.
Put the two together and a pattern emerges that matters more than either milestone on its own. Our Dallas–San Antonio milestone shows the Waabi Driver generalizes across where it drives. Zero-shot transfer of the Waabi Driver to the Volvo VNL Autonomous shows it generalizes across what it drives. Neither is a one-off trick — they're two demonstrations of the same underlying capability: an AI system that reasons about driving rather than one that has simply memorized it.
Since 2021, Waabi has worked at the frontier of Physical AI. This is exactly what that frontier looks like: intelligence that generalizes, safely, wherever it's needed next.