2026-06-18 — views

AV Consumer Adoption & Rider Experience Index — The Demand Side of the Ramp

Rider NPS, pricing vs Uber/Lyft, adoption curves, and whether real consumer demand supports AV scale targets — the demand side of the ramp.

Article 15 in the Physical AI Benchmark Series

Articles 10 through 14 dissected supply-side constraints: HD mapping bottlenecks, teleop staffing walls, OTA velocity, FMVSS regulatory gates, and capital structure. This article asks the question those articles deferred — is consumer demand actually there to justify the scale targets? The short answer is yes. The longer answer explains why that matters less than most people think.

Section 1 — Rider Experience Comparison

The most important benchmark in consumer AV is not safety or price in isolation — it is the combination of reliability, wait time, and price that determines whether a rider switches from Uber to an autonomous vehicle and stays switched.

Dimension	Waymo One	Tesla Robotaxi (Austin, est.)	Uber/Lyft (human driver)
Average wait time	3–6 min (Phoenix/SF mature markets)	Unknown (early launch)	3–8 min (varies by city)
Reported NPS (est.)	80+ (industry reports citing Waymo)	N/A — pre-data	~30–45 (Uber/Lyft typical, est.)
Price per mile	~$1.50–2.50	~$0.75–1.50 (stated target)	~$1.20–2.00
In-app experience	Waymo One app + Google Maps integration	Tesla app	Uber/Lyft apps
Safety perception (rider surveys, est.)	High — no fatalities at commercial scale	Unknown — new deployment	Medium — driver quality varies
Rider novelty factor	High (first-time wow effect)	Not yet established	None
Cancellation/reliability	Low cancellation (vehicles always available)	Unknown	Higher — driver-initiated cancels common
Accessibility (wheelchair etc.)	Limited (Jaguar I-PACE platform)	Unknown	Varies (UberAssist program)

NPS and safety perception figures are based on industry estimates and third-party rider surveys. Waymo has not publicly disclosed its exact NPS figure.

An estimated NPS of 80+ for Waymo — if accurate — would place it among the highest-rated consumer services in any sector. For context, Apple retail scores roughly 75–80 and Amazon sits around 65–70 in published estimates. Uber and Lyft’s typical range of 30–45 reflects the inconsistency of human driver quality. The Waymo advantage here is structural: no driver cancellations, no variable driver behavior, no social friction.

Section 2 — Adoption Curve Analysis

Consumer AV adoption follows a three-phase arc. Each phase has different price sensitivity, different barriers, and a different customer profile.

Phase 1 — Early Adopters (current, 2024–2026)

Tech-enthusiasts, tourists, and commuters in geofenced zones. This group pays a premium for novelty and consistency. Price sensitivity is low. Estimated 200K–500K total unique riders across all Waymo markets as of mid-2026 (est. — Waymo has not disclosed cumulative unique riders).

Defining characteristic: these riders chose Waymo because it is new and interesting, not because it is the cheapest or most available option.

Phase 2 — Pragmatic Adopters (2026–2029, est.)

Value-driven commuters who switch when AV price is at or below Uber AND reliability is proven through personal experience or social proof. Price sensitivity is high. Estimated requirements for Phase 2 conversion:

Pricing at or below comparable Uber ride
Sub-5-minute wait times in dense zones
99.9%+ trip completion rate (driver cancellations eliminated)
Geofence expansion to cover their actual travel patterns

This phase is where the volume numbers move. If Tesla hits its stated Cybercab target of $0.75–1.50/mile, Phase 2 activation happens earlier and faster than current models project.

Phase 3 — Mass Market (2029+, est.)

Non-car-owners, elderly populations, and mobility-impaired riders who gain new mobility options that did not exist before autonomous vehicles. This is not market share from Uber — it is total addressable market expansion. A 78-year-old who stopped driving five years ago is not currently an Uber user. They are a new rider.

This is where the true long-term scale opportunity lives. It is also the least modeled by current analyst consensus, which focuses on ride-hail share rather than mobility TAM expansion.

Adoption Barrier Ranking

The following table ranks consumer adoption barriers by stated concern in rider surveys. All percentages are estimates based on published AV readiness research (KPMG, AAA, J.D. Power); exact figures vary by survey methodology and year.

Barrier	% citing as concern (est.)	Waymo status	Tesla status
Limited geography/geofence	~52%	Major barrier — 4 cities only	Single city (Austin)
Safety concerns	~45%	Largely addressed — no fatalities at commercial scale	Unknown — new deployment
Price vs. Uber	~38%	Currently at parity or premium in most markets	Targeting below Uber
Comfort with no driver	~28%	Fading significantly with repeat use	Unknown
App/booking experience	~15%	Good — Google Maps integration reduces friction	TBD

The geofence barrier is the largest single adoption constraint — larger than safety. More than half of prospective AV riders cite “it doesn’t go where I need to go” as their primary barrier. This maps directly to the HD mapping bottleneck documented in article 10: Waymo’s 12–18 month pre-mapping requirement per city is not just a supply-side cost problem, it is directly suppressing demand conversion.

Section 3 — Waymo’s Demand Metrics (What Is Known)

Waymo does not publish detailed demand metrics. The following data points are drawn from public statements, blog posts, and third-party reporting:

150K+ weekly rides (mid-2026) — the fastest growth rate of any commercial AV operator globally, per public reporting
Waitlist demand — Waymo has reported waitlists in new markets before launch, indicating demand consistently exceeds fleet supply
Repeat rider rate — Waymo has cited high repeat usage (exact figures not disclosed) as evidence that satisfaction drives retention rather than novelty alone
Google Maps integration — Waymo One is bookable directly from Google Maps in San Francisco and Phoenix, materially reducing acquisition friction by placing the booking flow inside a product 2B+ users already open for navigation
Uber partnership (San Francisco) — Uber’s app surfaces Waymo as a vehicle option in SF, bringing AV access to Uber’s existing 150M+ active user base without requiring a new app download

The waitlist data point is particularly significant. In a normal ride-hail market, oversupply is the chronic problem — too many drivers relative to demand. Waymo’s waitlists invert this: too many riders relative to fleet. This is not a demand problem.

Section 4 — Price Competitiveness

The following table estimates the cost of a typical 5-mile urban trip across competing services. All figures are estimates based on public pricing, reporter accounts, and industry analysis. Actual prices vary by time of day, surge conditions, and city.

Service	Price (est.)	Wait time	Notes
Waymo One (San Francisco)	~$12–18	3–6 min	Premium pricing in early, capital-intensive market
Waymo One (Phoenix)	~$8–14	3–5 min	Larger geofence, more competitive pricing
Uber (national avg)	~$10–16	4–8 min	Driver supply-dependent; surges frequently
Lyft (national avg)	~$9–15	4–8 min	Similar structure to Uber
Tesla Robotaxi (target)	~$4–8	Unknown	If Cybercab hits $0.75–1.50/mile stated target

Waymo’s San Francisco pricing is currently at or above Uber for a typical trip. The Phoenix market, with a larger geofence and more mature operations, is more competitive. If Tesla achieves its stated $0.75–1.50/mile economics — roughly 40–60% below current Waymo pricing — it would be the cheapest motorized transport option available in any city where it operates, potentially triggering Phase 2 adoption earlier than current adoption models project.

The critical caveat: Tesla’s pricing target is stated intent, not demonstrated economics. The per-mile cost depends on Cybercab unit economics that have not been validated at commercial scale. Article 8 in this series covered the economics in detail. The gap between stated target and demonstrated cost remains an open variable.

Section 5 — The Demand Risk: Is Supply the Real Constraint?

The central finding of this demand-side analysis is that consumer demand is not the constraint on AV scaling. All available evidence points to supply as the binding variable:

Waymo operates with waitlists, not empty seats
Every new city launch is capacity-constrained (fleet size + mapping + teleop), not demand-constrained
The geofence barrier — the largest adoption concern in rider surveys — is itself a supply-side problem (HD mapping coverage)
Pricing is above the Phase 2 conversion threshold in the highest-profile market (SF), but below in mature markets (Phoenix) — and demand remains strong in both

The implication is structurally important for reading the Physical AI scorecard. The supply-side constraints documented in articles 10 through 13 — HD mapping, teleop staffing, OTA velocity, and the FMVSS gate — are not just cost and timeline problems. They are the mechanism by which a large, waiting demand pool is being rationed. Solving those constraints does not need to create demand. It only needs to serve demand that already exists.

This reframes the investment thesis. The question is not “will people ride in AVs?” — behavioral data already answers that. The question is “how fast can the operators build enough vehicles, map enough cities, and staff enough teleoperators to serve the riders who are already asking?”

The demand side of the ramp is not the ramp. It is the destination.

How This Article Fits the Series

This is article 15 in the Physical AI Benchmark Series. The series has now covered:

Articles 1–9: Technology, regulation, capital, and the master scorecard
Articles 10–13: Four supply-side structural constraints (HD mapping, teleop, OTA, FMVSS)
Article 14: Updated scorecard integrating all four constraints
Article 15 (this article): The demand side — rider experience, adoption curves, and pricing

The next article in the series will examine the humanoid ramp independently, applying the same supply-demand framework to Optimus and physical AI labor applications beyond ride-hail.