INFRA Signal 406
As Waymo expands its fleet from 700 in 2025 to nearly 4,000 now, recalls and edge-case failures like flooded roadways and a San Francisco power outage mount (Emmy Martin/New York Times)
↗
Waymo has grown its driverless fleet from 700 vehicles in 2025 to almost 4,000 today, and the larger fleet is encountering more recalls and failures in unusual conditions such as flooded streets and a citywide power outage.
The rapid scaling of autonomous-vehicle deployments is exposing software and hardware limits that were not evident at smaller volumes. Engineers must now allocate resources to detect, diagnose, and remediate edge-case scenarios that can interrupt service or trigger recalls. Operational reliability and safety compliance become more costly and complex as the fleet expands.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Fleet size has increased roughly sixfold, raising the absolute number of incidents that need to be managed.
Recent problems include vehicle behavior on flooded roadways and loss of functionality during a San Francisco power outage.
Addressing these issues will require expanded testing, monitoring infrastructure, and contingency handling for rare environmental conditions.
THE READ
What the cluster adds up to.
Waymo's expansion from a few hundred to several thousand autonomous cars means that any software defect or hardware flaw now affects a much larger customer base. The company has begun issuing recalls, indicating that the existing validation processes did not fully anticipate the scale of deployment. Engineers will need to augment their continuous-integration pipelines with broader scenario coverage to keep pace with the fleet growth. The reported failures highlight specific edge cases that were either under-tested or not considered in the original design. Flooded roadways can impair sensor perception, while a citywide power loss can disrupt communication and navigation services. Mitigating these problems will likely involve adding redundancy, improving sensor fusion algorithms, and building fallback modes that can operate with limited connectivity. Scaling the fleet also raises operational costs beyond the vehicles themselves. More frequent software updates, remote diagnostics, and field service interventions become necessary to address recalls promptly. Teams must budget for additional engineering headcount, testing environments, and data-collection infrastructure to support the
Written by elseif from the cluster below · checked for specifics the sources never containedTHE CLUSTER