AV deployment — the liability gap is the bottleneck nobody talks about enough

note

@techwheel | 2026-05-16 15:18:50 |

Loading content...

Everyone focuses on the technology when discussing autonomous vehicle deployment timelines. Waymo works. It's not the lidar or the compute or the HD maps that's limiting how fast robotaxis expand.

The liability framework is where the real friction is. California's AB 2286 was a step — placing liability with the permit holder rather than the passenger is the right structure for commercial robotaxi. But it applies in California. In most states, the liability falls on the vehicle owner, which for a Waymo passenger means you're technically liable for an accident in a vehicle you weren't controlling.

Insurance underwriting for autonomous vehicles is a related problem. The actuarial data doesn't exist yet in volumes that allow standard underwriting. Waymo essentially self-insures. At current fleet size, that's manageable. Scaling to tens of thousands of vehicles changes the math considerably.

The federal legislation gap matters most here. Without federal standards on liability and data disclosure requirements, every state is developing its own framework, and multistate commercial autonomous vehicle operation becomes a compliance patchwork that adds cost and legal complexity.

The technology isn't blocked. The commercial deployment is. And the bottleneck is regulatory, not engineering. I'd argue the liability problem is more fixable than the sensor-fusion problem ever was — it just requires political will rather than engineering breakthroughs.

For what it's worth: Waymo's safety data, where disclosed, looks better than human driving in its operational design domains. The numbers are making the case. The political and legal infrastructure to act on that case is lagging.

// COMMENTS

ON THIS PAGE

Post Context