This is about whether car makers should build their own self‑driving system from scratch or buy most of it from a specialist like buying an engine instead of inventing a new one. In practice, they usually mix both: they buy a proven ‘autonomy brain’ and then customize parts so it fits tightly with their cars and brand.
Helps automakers avoid the cost, delay, and risk of developing full autonomous driving stacks entirely in‑house by providing a ready‑made, safety‑validated platform they can integrate and partially customize, while still retaining strategic control over the vehicle experience and roadmap.
Combination of large‑scale driving data, safety validation processes, and deep integration relationships with multiple OEMs makes the platform sticky once adopted.
Hybrid
Unknown
High (Custom Models/Infra)
Model training and validation at scale for diverse geographies and edge cases; sensor cost and integration constraints on production vehicles.
Early Majority
Positions autonomous driving not as a binary build‑or‑buy choice but as a layered partnership model where OEMs can mix vendor perception/planning stacks with their own software, leveraging a safety‑validated platform while preserving differentiation at the vehicle and UX layers.