Waymo

This claim is circulating widely on X today, reportedly tied to TechCrunch reporting, but I couldn’t find confirmation of a broad Waymo freeway suspension across San Francisco, Los Angeles, Phoenix, and Miami specifically related to construction zones. What *is* confirmed: TechCrunch reported Waymo temporarily paused robotaxi operations in Atlanta (and discussed issues in San Antonio) after vehicles repeatedly encountered flooded roads. Earlier this month, Waymo voluntarily recalled \~3,800 vehicles for software updates tied to standing water detection on higher-speed roads. Construction zones, however, are a separate — and longstanding — challenge for the entire autonomous vehicle industry. Why construction zones are hard for AVs: Temporary lane shifts Inconsistent cones/barriers Human flaggers Poor or conflicting signage Constantly changing layouts These are classic “edge-case” environments where real-world variability breaks assumptions learned during training and mapping. Waymo has invested heavily in HD mapping, simulation, remote assistance, and iterative software updates, but messy urban construction remains one of the hardest operational problems for autonomy at scale. This fits a broader pattern across robotaxi deployments: Flooding / standing water Dark intersections during outages Emergency vehicles School buses Protests or vandalism Unpredictable human behavior Autonomous driving progress is increasingly less about whether the AI can drive under normal conditions, and more about how reliably it handles rare, chaotic, infrastructure-heavy edge cases. That said, Waymo is still the clear commercial leader in U.S. robotaxis, operating across roughly 10+ markets and serving hundreds of thousands of weekly rides. The bigger takeaway: Robotaxis are no longer a “can it work?” story. They’re becoming a reliability, scaling, and operational resilience story.