Rosa Del Mar

Daily Brief

Issue 83 2026-03-24

Commercial Scale And Rollout Velocity

Issue 83 Edition 2026-03-24 7 min read
General
Sources: 1 • Confidence: Medium • Updated: 2026-04-11 19:43

Key takeaways

  • Waymo provides over 500,000 fully autonomous rides each week.
  • The Waymo Driver uses a 360-degree multi-sensor suite combining cameras, lidar, and radar.
  • Waymo's depot operations include cars autonomously returning for low-energy or mess events, manual cleaning when flagged, and manual plug-in charging today.
  • Waymo's sixth-generation hardware stack keeps camera-radar-lidar modalities but is simplified and significantly lower cost, with sensor-plus-compute hardware claimed to be comparable to a high-end driver-assistance system.
  • Real-time driving inference for Waymo runs on-vehicle, with cloud connectivity used only for non-real-time auxiliary tasks.

Sections

Commercial Scale And Rollout Velocity

  • Waymo provides over 500,000 fully autonomous rides each week.
  • Waymo operates about 3,000 cars and drives over 4 million fully autonomous miles per week.
  • Waymo has fully autonomous operations in 11 U.S. cities, with public riders in 10 of them, and Nashville is a newly started non-public city.
  • Waymo's first fully autonomous commercial service began in 2020 in Chandler, Arizona using a fourth-generation system.
  • Waymo opened rider access in four new cities in a single day.

Multimodal Sensing And Joint Fusion Rationale

  • The Waymo Driver uses a 360-degree multi-sensor suite combining cameras, lidar, and radar.
  • Waymo describes lidar as providing high-resolution 3D sampling and radar as lower-resolution sensing that degrades better in adverse weather such as fog, snow, and heavy rain.
  • Waymo uses joint fusion of camera, lidar, and radar encodings rather than switching between sensors by environment.
  • Waymo still uses three sensing modalities (cameras, radar, and lidar) and has significantly optimized and simplified all three across generations.
  • Waymo detected and responded to a pedestrian occluded by a bus using a noisy signal from peripheral lidar reflections bouncing under the bus.

Operational And Physical World Bottlenecks

  • Waymo's depot operations include cars autonomously returning for low-energy or mess events, manual cleaning when flagged, and manual plug-in charging today.
  • Waymo optimizes autonomous driving not only for safety but also for smoothness and predictable, socially compatible driving behavior.
  • Waymo considers pickup and drop-off behavior a nuanced autonomy challenge involving rider intent, curb context, and minimizing disruption (e.g., avoiding blocking driveways or double-parking).
  • Cold winter weather is a weak point for generalization because it affects hardware needs such as sensor cleaning and heating elements and slippery-surface control.
  • Waymo riders are generally well-behaved and keep cars clean, though rider behavior varies by context such as a college town on a Saturday night.

Cost Down And Platform Portability Claims

  • Waymo's sixth-generation hardware stack keeps camera-radar-lidar modalities but is simplified and significantly lower cost, with sensor-plus-compute hardware claimed to be comparable to a high-end driver-assistance system.
  • Waymo states its driver software largely generalizes across vehicle platforms and sensor configurations.
  • Waymo attributes large declines in automotive radar costs to the shift from aviation-scale hardware to mass automotive supply chains, and says imaging radar remains costlier but is trending downward.
  • Waymo says lidar costs are following a predictable cost-down trend and that it is simplifying and optimizing lidar designs using learnings from prior generations.

Edge Plus Cloud Split And Offboard Ops Ml

  • Waymo's depot operations include cars autonomously returning for low-energy or mess events, manual cleaning when flagged, and manual plug-in charging today.
  • Real-time driving inference for Waymo runs on-vehicle, with cloud connectivity used only for non-real-time auxiliary tasks.
  • Waymo uses off-board models for post-ride functions such as detecting cleanliness issues or forgotten items and routing vehicles accordingly.

Watchlist

  • Waymo's sixth-generation 'Ojai' platform is a custom passenger-oriented vehicle planned to begin rolling out publicly this year.
  • Waymo plans to start operating in London and Tokyo this year and does not expect to deploy the San Francisco driver there without additional data collection, specialization, and validation.

Unknowns

  • What is Waymo's utilization per vehicle (rides per vehicle per day) and how does it vary by city, service hours, and service-area size?
  • What are the safety and reliability metrics (e.g., incident rates, intervention rates, downtime) associated with the reported scale, and how are they trending over time?
  • What is the actual sixth-generation hardware bill of materials and fully loaded capex per vehicle (including maintenance, spares, and depreciation), and how does it compare to prior generations?
  • What concrete evidence supports the claim that foundational world models are simplifying the autonomy system and improving scalability (e.g., fewer modules, fewer rules, shorter validation cycles)?
  • How complete is the 'ODD-based' portability in practice—what fraction of a new city launch is new data collection, new validation, new mapping/localization, and operational setup?

Investor overlay

Read-throughs

  • Waymo may be transitioning from pilot economics to repeatable commercial operations, as implied by over 500000 weekly rides and rapid multi city access expansion, shifting investor focus toward scaling playbooks and throughput constraints rather than core autonomy alone.
  • Sixth generation hardware simplification with camera radar lidar retained and sensor plus compute claimed comparable to high end ADAS could indicate a step down in per vehicle capital intensity, potentially improving unit economics if utilization and uptime hold.
  • Edge only real time inference with cloud used for auxiliary fleet and depot tasks suggests differentiation could come from an operations software stack that drives vehicle availability, turnaround, and rider experience, not just the driving model.

What would confirm

  • Disclosure of utilization by city such as rides per vehicle per day, service hours, and service area, showing stable or improving utilization as new cities and the Ojai platform roll out publicly.
  • Reported safety and reliability trends at current scale such as incident rates, intervention rates, and downtime, demonstrating improvement or stability as rollout cadence increases and new geographies launch.
  • Transparent sixth generation cost evidence such as hardware bill of materials and fully loaded capex per vehicle including maintenance, spares, and depreciation, showing meaningful reduction versus prior generations while maintaining multimodal sensing.

What would kill

  • Utilization per vehicle proves low or declines materially with city expansion, implying demand, service area, or operational bottlenecks prevent scalable revenue per asset despite high aggregate ride counts.
  • Safety or reliability metrics worsen with scale or during new city launches, indicating validation, winterization, pickup drop off, or depot constraints are gating rollout and increasing downtime.
  • Sixth generation hardware or operations costs are not materially lower when fully loaded, or require high manual depot labor such as cleaning and charging that offsets any sensor and compute savings.

Sources