Waymo’s robotaxis have moved from science fiction to street‑legal fleets, yet their promise to ease congestion remains unproven. Over a 1,000‑day span in California, the autonomous fleet logged 13.8 million rides and 86.3 million miles, but only 36 percent of those miles carried passengers at launch and the fleet had to rely on human drivers during peak.

Initial data showed only 36 percent of Waymo’s miles were passenger‑laden, rising to 56 percent by study end—a plateau that still leaves 44 percent of trips dead‑heading. These empty runs include idle vehicles waiting for fares and vehicles shuttling to pick‑up points, both of which dilute the system’s traffic‑reduction claims.

Waymo’s gradual shift to freeway service has trimmed dead‑head mileage per trip, but the company still operates 44 percent empty miles. A 2024 MIT Transit Lab study confirms that robotaxis match ride‑hail traffic levels, undermining the narrative that autonomous fleets will automatically unclog streets.

The persistence of dead‑heading indicates that autonomous fleets still require significant human intervention and infrastructure support. Until these inefficiencies drop, robotaxis will not deliver on their promise of smoother traffic flows, and their safety and cost advantages remain contingent on continued operational refinement.