Waive integrates Vision-Language-Action models (VLAs) to create a conversational interface for the car. This allows users to talk to the AI chauffeur ("drive faster") and provides engineers with a powerful introspection tool to ask the system why it made a certain decision, demystifying its reasoning.
Instead of building its own capital-intensive robotaxi fleet, Waive's go-to-market strategy is to sell its autonomous driving stack to major auto manufacturers. This software-centric approach allows them to leverage the scale, distribution, and hardware infrastructure of established OEMs to reach millions of consumers.
A key learning from working with auto manufacturers is the desire for brand differentiation through driving personality. Waive can tailor its AI's behavior—from "helpfully assertive" to comfortably cautious—to match a brand's specific identity. This transforms the AI from a utility into a core part of the product experience.
Waive treats the sensor debate as a distraction. Their goal is to build an AI flexible enough to work with any configuration—camera-only, camera-radar, or multi-sensor. This pragmatism allows them to adapt their software to different OEM partners and vehicle price points without being locked into a single hardware ideology.
Waive's core strategy is generalization. By training a single, large AI on diverse global data, vehicles, and sensor sets, they can adapt to new cars and countries in months, not years. This avoids the AV 1.0 pitfall of building bespoke, infrastructure-heavy solutions for each new market.
The AI's ability to handle novel situations isn't just an emergent property of scale. Waive actively trains "world models," which are internal generative simulators. This enables the AI to reason about what might happen next, leading to sophisticated behaviors like nudging into intersections or slowing in fog.
The key difference between AV 1.0 and AV 2.0 isn't just using deep learning. Many legacy systems use DL for individual components like perception. The revolutionary AV 2.0 approach replaces the entire modular stack and its hand-coded interfaces with one unified, data-driven neural network.