2026-06-18 👁️ — views

Physical AI Accessibility — How AVs Could Transform Mobility for 70M+ Non-Drivers

How autonomous vehicles could unlock independent mobility for the 70M+ Americans who cannot legally or safely drive — the market and regulatory case.

Article 44 in the Physical AI Benchmark Series — The Accessibility Dimension

Every article in this series so far has framed the autonomous vehicle ramp as a convenience upgrade: people who already drive gaining time back, reducing stress on long commutes, or hailing a robotaxi instead of parking in a dense city. That framing describes one market. It overlooks a larger one.

Roughly 1 in 5 Americans cannot legally or safely operate a motor vehicle. That is more than 60 million people — potentially more than 70 million when all overlapping populations are counted — for whom a car is not a convenience tool but an inaccessible one. For these individuals, autonomous vehicle technology is not a quality-of-life upgrade. It is a potential first-time mobility solution at a scale that no previous technology has offered.

This article examines the non-driver population, the inadequacy of current alternatives, what AV accessibility looks like in practice today, why the accessibility case creates political capital that accelerates regulatory approval, and why Waymo’s driverless operation already serves this population in a way Tesla’s supervised FSD cannot.

Section 1 — The Non-Driver Population

The AV industry is typically discussed as a product for people who already drive. The larger opportunity is replacing car dependency for people who cannot drive at all.

The totals overlap significantly — a person with epilepsy may also be over 70 — so the unduplicated count is lower than the column sum. Even with conservative deduplication, the non-driver population represents roughly one-fifth of the US total. This is not a niche accessibility market. It is a structural gap in the transportation system.

Several features of this population are worth noting. First, it is growing: the US population over 65 is projected to double by 2060 (est.), and driving cessation typically occurs in the 70s when vision and reaction time decline below safe-driving thresholds. Second, the gap is geographically concentrated in auto-dependent suburbs and rural areas where transit is minimal — exactly the environments where car access is most critical for basic errands like grocery shopping and medical appointments. Third, the population includes groups with significant political voice: elderly voters turn out at above-average rates, and disability advocacy organizations have a long track record of successful regulatory pressure.

Section 2 — Why Current Alternatives Are Inadequate

The non-driver population has not been without mobility options. The problem is that every existing option fails on at least two of four critical dimensions: coverage, cost, reliability, and dignity.

Paratransit is the government’s answer to non-driver mobility and it is legally mandated under the Americans with Disabilities Act for transit agencies operating fixed-route service. In practice, it is expensive for agencies (see Section 4), constrained to urban service areas, requires advance booking, and frequently fails on reliability. A non-driver in a suburban or rural area has no ADA paratransit right at all — the mandate only applies where fixed-route transit exists.

Ride-hail covers more geography and does not require advance booking, but wheelchair-accessible vehicles are rare on both Uber and Lyft platforms, supply is inconsistent, and the networks depend on human drivers willing to serve accessibility trips. Family dependency — a parent, spouse, or adult child serving as unpaid chauffeur — is the de facto solution for millions of Americans, but it transfers a significant time and schedule burden onto caregivers who are often working-age family members with their own constraints.

The gap is specific: no current solution provides on-demand, affordable, wheelchair-accessible, independent mobility at scale outside of major metro areas. That is the gap autonomous vehicles are positioned to close.

Section 3 — What AV Accessibility Looks Like in Practice

Waymo’s Accessibility Record

Waymo One is ADA-compliant: the Jaguar I-PACE vehicle fleet and the newer Gen 6 platform both accommodate folding wheelchairs and common assistive devices. More significantly, Waymo has run pilot programs with the National Federation of the Blind — blind users have completed solo Waymo rides without any sighted assistance present in the vehicle.

The key feature that makes Waymo’s service genuinely accessible, rather than nominally accessible, is driverless operation. A blind rider using Waymo in Phoenix does not need a sighted human present to manage communication with a driver, explain their destination, or navigate an in-vehicle interface designed for sighted users. The voice interface handles the entire trip interaction. The vehicle arrives, confirms the pickup, and delivers the passenger to the destination — no human driver mediation required.

Waymo’s single largest deployment city, Phoenix, is well-suited to this use case. The Phoenix metro area — including Chandler and Scottsdale — has a significant and growing retirement community population. The city’s suburban, wide-street layout is also the operating environment where Waymo’s mapping-intensive approach works best and where transit alternatives are thinnest.

Tesla’s Accessibility Path

Tesla’s current FSD product is supervised: it legally requires a licensed, attentive driver in the seat. By definition, supervised FSD is not accessible to people who cannot drive. The technology may assist a licensed driver in reducing workload, but it does not serve the non-driver population.

The Cybercab — Tesla’s planned driverless two-seat vehicle with no steering wheel or pedals — is potentially accessible to non-drivers, but it requires a Federal Motor Vehicle Safety Standards exemption for pedal-free operation. That regulatory hurdle has not yet been cleared as of mid-2026 (est.). When and if the Cybercab reaches commercial driverless operation, it would enter the accessibility market. Until then, Tesla’s products do not serve non-drivers.

Tesla’s Optimus humanoid robot could eventually assist mobility-impaired individuals in home settings — helping with transfers, retrieving objects, providing physical assistance with daily tasks. But commercial personal-assistance deployment is years away (est.) and addresses a different dimension of disability support than transportation independence.

Section 4 — The Regulatory Motivation: Accessibility as Political Capital

A recurring question in the AV policy conversation is why cities approve AV operating permits despite public safety uncertainty. The accessibility case is one significant part of the answer.

Paratransit cost pressure. US transit agencies spend an estimated $5 billion per year on paratransit — a legally mandated, operationally expensive service that agencies are required to provide but that strains transit budgets disproportionately. A typical ADA paratransit trip costs transit agencies far more per trip than a fixed-route bus ride covering the same distance (est.). On-demand AV paratransit — vehicles that can be dispatched without human driver labor cost, serve any address, and complete trips without the scheduling inefficiency of shared-ride routing — could reduce per-trip costs by 30–60% (est.) at equivalent or better service levels. Agencies facing budget pressure have a direct financial motivation to see AV paratransit technology succeed.

Political constituency. The elderly vote at above-average rates. Disability advocacy communities — the National Federation of the Blind, the National Council on Disability, state Independent Living centers — have organized political voices and long histories of successful regulatory advocacy. AV accessibility is one issue where these communities are aligned with AV deployment, creating a political counterweight to the safety-skeptical voices that dominate public AV discourse.

Federal funding signals. The US Department of Transportation and the National Highway Traffic Safety Administration have explicitly included transportation accessibility as a policy goal in AV regulatory frameworks. NHTSA’s automated vehicle guidance identifies access for people with disabilities as a core public benefit that AV technology should serve. Federal grant programs for AV pilots have included accessibility as a scoring criterion.

Rural mobility crisis. Approximately 20% of Americans live in areas with no fixed-route transit service and limited or no Uber and Lyft coverage. Human-driven ride-hail requires driver supply density that rural areas cannot support. An AV on-demand service can serve rural routes at a cost point that human ride-hail cannot, and can do so without the wait times and availability gaps that make current rural ride-hail unreliable. Rural accessibility may prove to be the first AV deployment context that achieves economic viability at scale ahead of urban robotaxi — because the alternative is nothing, not a slower human-driven substitute.

Section 5 — Accessibility as a Ramp Accelerator

The accessibility market changes the AV ramp calculus in two distinct ways that compound over time.

The Demand-Side Case

Accessibility creates a captive market that will pay — or whose government will pay — for AV service even when prices remain above human-driven ride-hail rates. A paratransit AV trip that replaces a $60 government paratransit ride is economically viable for the transit agency at a vehicle operating cost substantially above an Uber fare. This is not a subsidy-dependent business model in the long run; it is a pricing structure where the incumbent alternative is already expensive. AV operators do not need to undercut human-driven ride-hail to win this market. They need to be more reliable, more available, and less expensive than the government-operated service they are replacing.

The Regulatory-Side Case

Accessibility use cases may receive faster or more favorable regulatory treatment than general-public robotaxi precisely because the counterfactual for the non-driver passenger is clearly worse. A blind rider choosing Waymo driverless is making an informed consent decision with no adequate alternative. A city reviewing that rider’s situation weighs AV risk against the alternative: no independent mobility at all. That calculus is different from a sighted rider choosing AV over a human-driven Uber. Regulators reviewing accessibility applications are comparing AV risk to a baseline of immobility, not a baseline of adequately-served transportation. That framing favors approval.

Waymo’s Structural Head Start

Waymo’s driverless operation — no safety driver present in the vehicle — is the necessary condition for genuine non-driver accessibility. A Waymo ride with a safety driver seated in the vehicle is still nominally accessible, but the safety driver’s presence complicates the experience for blind riders and does not fully realize the independence that driverless operation enables.

Tesla’s supervised FSD requires a licensed driver in the seat by design and by law. Tesla does not currently offer a driverless service. This means Waymo already serves the non-driver accessibility population in a way Tesla cannot replicate until the Cybercab achieves federal regulatory approval for driverless pedal-free operation.

This is not a minor competitive distinction. It means Waymo has a first-mover advantage in the accessibility market that is not a function of technology alone — it is a function of regulatory approval for full driverless operation that Tesla has not yet obtained for any platform. In the accessibility context, “first to driverless” is more valuable than “first to high-capability supervised FSD,” because supervised FSD does not serve the non-driver population at all.

The accessibility dimension of the AV ramp is not a footnote to the convenience story. It is a parallel market with distinct economics (government paratransit replacement), distinct regulatory dynamics (immobility baseline changes the risk calculus), and a distinct competitive structure (driverless-only products serve a population that supervised-FSD products cannot reach). Waymo’s current Phoenix deployment — driverless, ADA-compliant, voice-interface accessible — already demonstrates this market in early form. Whether AV operators recognize the accessibility market as a strategic priority, rather than a compliance byproduct, will shape where the next city expansion decisions land.

Sources: Waymo blog (waymo.com/blog); USDOT FTA ADA paratransit guidance (transit.dot.gov); NHTSA Automated Vehicles for Safety (nhtsa.gov); CDC Disability and Health data (cdc.gov/ncbddd/disabilityandhealth). All figures marked (est.) are estimates based on public disclosures, government reports, and third-party research; they have not been independently verified and may differ from primary source data.

Sources

• Waymo + National Federation of the Blind partnership — Waymo blog ↗
• ADA paratransit requirements — USDOT FTA ↗
• NHTSA Automated Vehicles for Safety — NHTSA ↗
• CDC disability statistics — US Centers for Disease Control ↗

Tags