2026-06-18 👁️ — views

AV Accessibility — Independence for 65 Million Americans Who Cannot Drive

For 65 million non-driving Americans — elderly, disabled, vision-impaired — autonomous vehicles are not a convenience. They are independence.

Article 93 in the Physical AI Benchmark Series — AV Accessibility: The 65 Million Americans Who Cannot Drive and Why Autonomous Vehicles Are Their Most Important Technology in a Generation

Autonomous vehicles are discussed almost exclusively as a convenience technology for people who already have transportation — a way to commute hands-free, summon a robotaxi, or reduce the friction of urban travel. The story that receives far less attention is the inverse: for approximately 65 million Americans who cannot or should not drive, AV is not a convenience upgrade. It is the first credible technology in decades that can restore independent mobility.

This population includes elderly adults who have surrendered their licenses due to age-related decline, people with visual impairments who have never been able to obtain a standard driver’s license, adults with physical disabilities that prevent safe vehicle operation, people with cognitive disabilities for whom the cognitive load of driving exceeds safe capability, and adults with medical conditions — such as epilepsy — that impose driving restrictions. For these groups, the current transportation alternatives are systematically inadequate: expensive, slow, advance-scheduled, geographically limited, or dependent on caregiver availability.

This article maps the accessibility dimension as a Physical AI benchmark: who the affected population is, how large the mobility gap is, what AV delivers differently, and what the early deployment evidence from Waymo shows.

Section 1 — The Scale of Mobility Limitation in the US

The non-driving population is large, structurally underserved, and growing as the US population ages.

The 65 million figure (est.) encompasses overlapping categories — not every person with a cognitive disability also has a physical disability — but the aggregate non-driving population is one of the largest underserved markets in US consumer mobility.

The Current Alternatives Are Systematically Inadequate

The transportation options available to non-drivers today share a common deficiency: none provides the combination of independence, on-demand availability, and geographic coverage that personal vehicle access delivers.

Fixed-route public transit covers approximately 30% of US land area (est.) and is structurally inaccessible for many people with mobility limitations — buses and subways require physical ability to board, navigate, and manage connections. Service does not extend to residential front doors. Rural and suburban areas where much of the elderly and disabled population lives are largely excluded.

Paratransit — the ADA-mandated demand-responsive service for people with disabilities — addresses some of these gaps but creates new ones. ADA paratransit requires 24-72 hours advance booking, operates at $30-50+/trip (est.) in many markets, is geographically restricted to corridors near fixed transit routes, and is chronically underfunded. Spontaneous transportation — a medical emergency at midnight, an unplanned social visit — is not available under the paratransit model.

Family and caregiver transport provides independence of a kind, but it is not independence — it is dependence transferred from a vehicle to a person. Caregiver availability is not guaranteed, imposes scheduling constraints on the caregiver’s own life, and disappears when the caregiver moves, becomes ill, or is simply unavailable. Many elderly adults explicitly report that losing the ability to drive — not the driving itself, but the dependency that follows — is the most significant quality-of-life reduction in late age.

Current ride-hail platforms (Uber, Lyft) create their own barriers for this population. The social dynamic of an unknown human driver is a significant deterrent for riders on the autism spectrum, those with social anxiety, women traveling alone at night, and others for whom unpredictable human interaction is a friction or safety concern. ADA-accessible vehicles on ride-hail platforms are inconsistently available — requests for wheelchair-accessible vehicles frequently fail. And people with visual impairments or significant cognitive limitations may not be able to use a smartphone app interface reliably or independently.

Section 2 — What AV Delivers Differently

The core value proposition of AV for the non-driving population is not convenience — it is independence. The specific capabilities that AV delivers differently from any current alternative:

The phrase “fundamental capability unlock” deserves emphasis. For someone who has never been able to drive — a person with severe visual impairment, for example — or for someone who has lost the ability after years of independence, AV does not provide a marginal improvement over the status quo. It provides a category-level restoration of capability that has not previously existed in a cost-accessible, on-demand, 24/7 form.

An elderly adult who surrenders their license at age 75 in a suburban or rural area loses the ability to drive to medical appointments, grocery stores, social activities, and family visits without depending on others to schedule around them. AV restores that independence without requiring the person to acquire new skills, purchase new equipment, or wait for a caregiver’s availability. The summon interaction can be as simple as a voice command to a smart speaker.

Section 3 — Waymo’s Accessibility Evidence

Waymo’s commercial deployment in Phoenix and San Francisco provides the most concrete early evidence of how AV is being used by the non-driving population. The data is largely qualitative and drawn from Waymo’s own public statements and press coverage, but the patterns are consistent and directionally significant.

The Phoenix market is informative precisely because it is not a dense urban core. Phoenix’s suburban sprawl — with large retirement communities in the metro area, low transit coverage, and car-dependent geography — is a closer analog to where much of the US elderly and disabled population actually lives than San Francisco’s dense urban grid. The fact that elderly riders are taking Waymo daily for errands and appointments in Phoenix suggests the use case is real and persistent, not a novelty.

The vision-impaired early adopter pattern is also significant. Blind and low-vision individuals have historically been among the most underserved by transportation alternatives. They cannot hail a physical taxi reliably, app-based ride-hail requires visual or high-cognitive-load navigation of a smartphone interface, and paratransit is advance-scheduled and geographically limited. Waymo’s model — summonable without visual verification, no driver social dynamic, interior-announced arrival — addresses the specific barriers this population faces.

Section 4 — The Healthcare Transport Opportunity

Non-emergency medical transportation (NEMT) represents one of the most commercially compelling AV accessibility use cases, because it combines a large, recurring, Medicaid-reimbursable demand with the specific operational characteristics that AV delivers better than current alternatives.

The NEMT case is structurally different from general ride-hail for accessibility purposes. The rider population is identified in advance — Medicaid beneficiaries with scheduled medical appointments — which allows fleet operators to pre-position vehicles, optimize routes, and integrate directly with healthcare scheduling systems. The revenue is Medicaid-reimbursable and therefore relatively insulated from the demand volatility that affects general consumer ride-hail.

The chronic problem in current NEMT is driver-side reliability. Contracted NEMT drivers have high turnover, inconsistent vehicle quality, and no-show rates that cascade directly into missed medical appointments. Dialysis patients who miss a treatment because their ride did not appear face acute medical consequences. Chemotherapy patients with compromised immune systems cannot use crowded public transit as a fallback. The AV proposition for NEMT is not just cost reduction — it is reliability as a medical-grade service characteristic.

Waymo’s early pilots with medical transport partnerships in Phoenix are the leading indicator that this use case is being taken seriously by at least one commercial AV operator. The commercial structure — direct healthcare system contracting rather than consumer ride-hail — is a distinct revenue model that may develop in parallel to the general consumer robotaxi business.

Section 5 — Policy and Regulatory Dimensions

The deployment of AV for accessibility populations raises policy questions that current regulation has not fully resolved.

The ADA compliance question is the most immediately consequential. If commercial AV operators are classified as public accommodations under ADA Title III, they will be required to ensure that a meaningful portion of their fleet is wheelchair-accessible and that the passenger interface is accessible to people with visual and cognitive impairments. This is an operational and design requirement that must be built into fleet composition and software from the beginning — it cannot be easily retrofitted.

The Medicaid NEMT regulatory gap is also significant. CMS’s silence on AV-specific NEMT guidance creates uncertainty about reimbursement structures, liability standards, and the documentation requirements for AV operators to qualify as NEMT providers. State Medicaid programs are not uniform in how they contract with NEMT providers, and the question of whether an autonomous fleet operator qualifies under existing NEMT provider certification frameworks is unresolved in most states.

Section 6 — The Accessibility Case as a Physical AI Benchmark Dimension

Framing AV accessibility as a Physical AI benchmark dimension — rather than as a social policy issue — has a specific analytical purpose. Market size, commercial viability, and deployment evidence for non-discretionary use cases provide a floor beneath the Physical AI ramp that is independent of consumer adoption enthusiasm.

The non-driving population represents demand that is:

1. Large — 65 million (est.) Americans with genuine mobility gaps
2. Non-discretionary — medical appointments, grocery access, and social connection are not optional activities
3. Underserved by alternatives — no current solution adequately addresses the combination of independence, on-demand access, 24/7 availability, and affordability
4. Commercially reimbursable — NEMT is Medicaid-reimbursable, providing a revenue stream independent of consumer willingness-to-pay for convenience
5. Geographically distributed — concentrated in suburban and rural markets where transit coverage is lowest and AV’s non-driver value proposition is strongest

This combination creates a fundamentally different demand curve for AV in accessibility use cases than in general consumer ride-hail. Consumer ride-hail adoption is highly elastic with respect to convenience and price — it competes with personal vehicle ownership, which is already available to most potential users. Accessibility AV competes with inadequate alternatives. The switching cost from paratransit to an AV service that is on-demand, 24/7, and similarly priced is essentially zero.

The implication for the Physical AI ramp is that the accessibility segment is likely to be adopted faster and with less price sensitivity than the general consumer segment — once deployments reach the geographic markets where the non-driving population is concentrated. That geographic expansion — from dense urban corridors to suburban and rural markets with lower lane complexity but higher accessibility demand — is the key deployment milestone to watch.

Section 7 — About This Series

This is article 93 in the Physical AI Benchmark Series. Previous articles have covered the ramp index, the humanoid race, unit economics, global competition, HD mapping, software and OTA updates, consumer demand, competitive moats, safety data, Waymo Gen 6, Optimus manufacturing, scorecard snapshots, 2030 forecast scenarios, the investor framework, city expansion pipelines, Tesla FSD state approval maps, AV weather and climate constraints, regulatory calendars, robotaxi fare pricing, humanoid deployment trackers, supply chain analysis, consumer adoption demand index, valuation and IPO analysis, the Physical AI 2026 mid-year roundup, AV unit economics cost-per-mile breakdown, the AV data flywheel comparison, the Physical AI supply chain, AV fleet operations, the full lifecycle environmental cost, the accessibility layer, the mapping architecture comparison, the China AV race, simulation and synthetic data training, AV urban planning and city impact, autonomous trucking freight economics, the European AV competitive landscape, the AV sensor technology debate, AV safety metrics, the AV talent war, the global AV regulatory map, AV financial sustainability burn rates, the Tesla Cybercab versus Waymo Gen 6 head-to-head (article 84), AV cybersecurity attack surfaces (article 85), the humanoid robots commercial deployment landscape (article 86), AV fleet electrification and the charging race (article 87), AV data as a business (article 88), AV insurance and liability (article 89), the driverless cabin and passenger experience (article 90), the Physical AI investment landscape (article 91), and AV safety vs human drivers statistics (article 92).

This article adds the accessibility dimension: the 65 million non-driving Americans, the inadequacy of current alternatives, what AV delivers differently, Waymo’s early deployment evidence, the NEMT healthcare opportunity, and the policy landscape.

Note: Population estimates, market sizes, and cost figures in this article are directional estimates drawn from CDC disability data, AAA Foundation older driver research, CMS NEMT program data, Waymo public statements, and press coverage as of mid-2026. Where data is uncertain or estimated, figures are labeled “(est.)” and should be treated as directional rather than confirmed definitive figures. This article does not constitute medical, legal, or policy advice.

Sources

• Waymo accessibility — Waymo blog ↗
• ADA transportation requirements — ADA.gov ↗
• Non-emergency medical transportation — CMS ↗
• Americans with disabilities data — CDC ↗
• Older adults and driving — AAA Foundation ↗

Tags