One tradeoff I think is still underrated is organizational learning velocity.

Build vs. buy isn't just about cost, differentiation, or time-to-market. Every capability you outsource is also a capability your engineering team doesn't develop internally. Over a 10–15 year vehicle lifecycle, that knowledge gap can become a bigger constraint than the original technical decision.

The strongest hybrid strategies don't just integrate vendor technology they deliberately retain ownership of the architectural decisions that determine how the entire system evolves over time. That's what keeps optionality alive as both AI and hardware continue to change.