The Global Expansion of Self-Driving Vehicles

What It Covers

Alex Wilhelm hosts Ben Seidel (AutoLane), Nathan Parker (Edge Case), and Ming Ma (Move) to examine the current state of autonomous vehicles across robotaxi expansion, fleet operations, safety certification, and autonomous commerce. The Uber-Zoox Las Vegas partnership announcement anchors a broader discussion of regulatory fragmentation, infrastructure bottlenecks, and the path to commercial scale in 2026.

Key Questions Answered

• •ODD Expansion Strategy: Autonomous vehicle companies expand city-by-city by solving only the "delta" between a cleared Operational Design Domain and the new target city. Moving from Las Vegas to Los Angeles requires less engineering work than moving to an entirely different climate zone because sensor calibration, traffic patterns, and road geometry overlap significantly — reducing re-certification time and cost for each geographically similar market.
• •Fleet Infrastructure Costs: Vehicle purchase represents only 30–40% of a five-year autonomous fleet total cost to serve. The remaining 60–70% covers charging infrastructure, maintenance facilities, insurance, cleaning, and parking. Fleet operators must provision 3–10 megawatts of power per facility, a process that takes 12–18 months with local utilities — making power provisioning one of the primary rate limiters to AV fleet scaling.
• •AV Fleet Management vs. Traditional Fleet: Managing an autonomous vehicle fleet requires airline-level safety protocols, not traditional ride-hail logistics. Every returning vehicle undergoes documented inspection with torque values recorded for each repaired component. Supply-demand curve optimization is handled algorithmically per city, and fleet operators maintain multiple facility types — charging, maintenance, and overflow parking — across each market they operate.
• •Safety Certification Scales Permanently: Edge Case Research's safety work does not end at launch. As fleets grow from 50 to 100,000+ vehicles, operators must monitor whether every unit still performs to the original certified safety standard. Component obsolescence, configuration changes, and hardware drift across large fleets require real-time observability platforms — modeled on airline fleet grounding protocols — to catch system-wide safety deviations before incidents occur.
• •Single Incident Risk: One Cruise-level safety incident involving Tesla or Waymo could pause the entire autonomous vehicle industry for 12–18 months. Public trust remains fragile enough that a San Francisco neighborhood mobilized a political campaign against Waymo over a single cat fatality. Operators, fleet managers, and safety firms all share downside risk from any high-profile incident regardless of which company causes it.
• •Autonomous Commerce Still Pre-Scale: Autonomous delivery — spanning sidewalk robots (Serve Robotics, Coco), drones (Wing, Zipline), and purpose-built vehicles (DoorDash's Fremont robot) — collectively holds negligible market share versus Uber Eats and DoorDash as of early 2026. AutoLane's thesis positions passenger cars as the dominant autonomous delivery form factor long-term, building an orchestration layer that routes orders across all vehicle types rather than betting on a single hardware format.