2026-06-18 👁️ — views

Physical AI Urban Design Impact — How AV Fleets Are Reshaping Parking, Curb Management, and City Infrastructure

AV fleets will strand 800M US parking spaces and make curb management a top city revenue source, reshaping urban land use for Tesla and Waymo.

Article 145 in the Physical AI Benchmark Series — Physical AI Urban Design Impact: How AV Fleets Are Reshaping City Infrastructure, Parking, Road Design, and the Future of Urban Planning

As AV fleets scale from thousands to millions of vehicles, they will fundamentally reshape how cities are built. Parking garages become stranded assets; road geometry needs redesign for vehicles that never park on-street; curb management becomes the critical urban resource. Tesla’s depot-free model vs Waymo’s depot model have very different urban footprints. This article maps the second-order urban infrastructure impact of Physical AI at scale.

All figures labeled “(est.)” are derived from public disclosures, industry research, urban planning literature, and analyst estimates rather than independently verified primary data. This article does not constitute investment advice.

Section 1 — The Parking Stranded Asset Problem

The Stranded Asset Cascade

The parking stranded asset problem unfolds in three phases. In the first phase (now through ~2028), AV deployment is too small to materially reduce parking demand. Individual cities begin updating parking minimum requirements; a few high-profile parking garages near Waymo operational zones see utilization decline at the margins.

In the second phase (~2028-2035), AV fleets in high-density operational zones reach the threshold where on-street parking demand measurably drops. Cities begin converting street parking lanes to bike lanes, bus lanes, and expanded sidewalks — a trend already underway in SF and LA for non-AV reasons. Parking garage operators begin factoring AV risk into 20-year bond financing. Cities that financed parking infrastructure via municipal bonds face stranded asset liability on their balance sheets.

In the third phase (~2035-2045), AV fleets approach mainstream adoption in major metros. The ~800M (est.) US parking spaces — built over 70 years at enormous public and private expense — face systematic underutilization. The most valuable conversion: structured parking garages in urban cores, where floor plates can be repurposed for housing, retail, or commercial uses once ramps and structural columns allow. The least valuable: surface lots in suburban areas, where land value is lower but conversion to other uses is easier.

The Sun Belt / dense-city asymmetry matters. Phoenix, Dallas, Houston — cities that built for the car — have excess surface parking that can be cheaply converted to AV staging areas or simply left idle as demand evaporates. NYC, SF, Boston — where parking is scarce and expensive — will see more violent repricing when AV demand compression hits supply that was already undersupplied.

Section 2 — Curb Management: The Critical Urban Resource

Why the Curb Is the New Oil

Urban planners are beginning to recognize that the curb is the scarcest and most contested public resource in a dense city — not road lane capacity, not parking supply, but the physical edge between the roadway and the building. Every use case competes for it: bike lanes, outdoor dining, loading zones, bus stops, and now AV PUDO. The curb cannot be widened; it is a fixed resource whose value increases with every new demand placed on it.

The shift from parking revenue to PUDO revenue is not obvious but is structurally significant. Today, cities monetize the curb primarily through parking meters and parking fines — a reactive, time-based revenue model that requires human enforcement. AV fleets create the conditions for a real-time, transaction-based curb pricing model: every PUDO event is a digital transaction that can be priced dynamically, metered to the second, and billed automatically to the AV fleet operator. At scale, this model could generate $500M-2B/year (est.) in new city revenue across US metros — potentially exceeding current parking meter revenue.

The equity dimension is the hardest policy challenge. Waymo’s city permit process in California includes service equity requirements — a condition that the company must demonstrate service in underserved neighborhoods, not just high-revenue corridors. Tesla’s owner-operator model has no equivalent mechanism: individual Cybercab owners will route to where demand and fares are highest, which in American cities means wealthy neighborhoods get service first. City governments that want equitable AV access will need to attach PUDO zone allocation to equity service requirements — a policy tool that doesn’t yet exist in most municipalities.

Section 3 — Road Design for AV Fleets

The Road Redesign Horizon

Road design changes on a 30-50 year cycle — the typical lifespan of road infrastructure before major reconstruction. This means road optimization for AV fleets is not a 5-year policy question; it is a 30-year structural shift embedded in city master plans, transportation improvement programs, and capital budgeting cycles.

The lane width opportunity is the most impactful near-term redesign lever. Standard 12-foot lane widths in US cities were designed with a 2-foot buffer for human error — drivers who drift, distracted, or misjudge lane position. AVs with centimeter-level lane-keeping precision do not need that buffer. Narrowing lanes from 12 to 9-10 feet (est.) on a standard 4-lane arterial road could add a 5th lane — a 25% throughput increase on the same right-of-way. That is the equivalent of adding a lane without acquiring new land. At city scale, this is a multi-billion-dollar infrastructure value creation that costs only the paint and signal reprogramming.

The V2X signal coordination opportunity is the most technology-dependent. Vehicle-to-infrastructure communication — where AVs communicate their speed, direction, and intent to traffic signals in real time — enables “green wave” optimization: a coordinated series of green lights that an AV fleet can ride through an arterial corridor without stopping. Modeled throughput gains from V2X signal coordination range from 15-40% on signalized corridors (est.), with additional fuel/energy savings from eliminated stop-and-go cycles. US DOT V2X infrastructure deployment is projected for 2027-2032 (est.) — slow enough that AV fleets will scale ahead of the supporting infrastructure in most cities.

The road marking sensitivity issue is a real near-term constraint for Tesla’s FSD. Camera-only perception relies on high-contrast lane markings that are often faded, obscured by snow or rain, or missing on older road sections. Waymo’s LiDAR + camera fusion is less sensitive to marking quality. This creates a geographic reliability gradient: FSD performs best on recently remarked highway lanes in well-maintained states and worst on old urban arterials in underfunded municipalities. Road marking quality is a $20-50B (est.) deferred maintenance problem across US cities — one that AV deployment will create pressure to address.

Section 4 — Waymo vs Tesla Urban Footprint Comparison

The Depot Model vs the Supercharger Model: A Strategic Comparison

Waymo’s depot model and Tesla’s Supercharger model represent two fundamentally different theories of how AV infrastructure integrates into urban fabric. Waymo’s approach is centralized and negotiated: secure a large site per city, coordinate with local government for permits and PUDO zones, build city-by-city relationships. Tesla’s approach is distributed and market-driven: use the existing 50,000+ (est.) Supercharger network globally as the backbone, let individual Cybercab owners charge anywhere, rely on market signals rather than city permits.

The depot model’s strategic advantage is relationship depth. Waymo’s per-city negotiation process — while slower and more expensive than Tesla’s distributed approach — builds city government partnerships that yield tangible operational advantages: priority PUDO zone placement, early warning on road construction that affects operational zones, and political cover when AV incidents require government support. In regulated markets (California most prominently), these relationships are not optional; they are the price of operating.

The Supercharger model’s strategic advantage is scalability. Tesla can enter a new city with essentially zero new infrastructure investment — the Supercharger network may already be present, and Cybercab owners use it commercially without any Tesla depot build-out. This makes Tesla’s geographic expansion cost curve dramatically lower than Waymo’s. In a winner-take-most AV market, first-mover advantage in new geographies matters enormously. Tesla’s ability to enter 50 US cities simultaneously versus Waymo’s sequential city-by-city build-out is a structural speed advantage.

The dead mall opportunity is a genuine strategic asset for Waymo. Dead malls — of which the US has ~1,000+ (est.) — typically feature large surface parking fields, industrial-zoned adjacencies, and highway access, all of which are optimal AV depot characteristics. A Waymo partnership with a mall owner or REIT to repurpose a dead mall parking field as an AV depot solves two problems simultaneously: Waymo gets cheap, large, highway-accessible depot space; the mall owner gets a tenanted use for otherwise worthless asphalt.

Section 5 — Urban Design Benchmark Scorecard

The Long Arc: AV Fleets as the Biggest Urban Design Force Since the Interstate Highway Act

The Interstate Highway Act of 1956 was the last time the United States made a technology decision that fundamentally reshaped every American city. The highways that followed didn’t just move cars faster — they determined where suburbs grew, which urban neighborhoods were bisected and hollowed out, where commercial development clustered, and what the American city looked like for the next 70 years.

AV fleet deployment at scale is the next technology decision of that magnitude. The difference is that the Interstate Highway Act was a deliberate top-down policy choice by the federal government. AV urban redesign will happen bottom-up, city by city, company by company, through thousands of permit decisions, curb management policies, parking minimum reforms, and V2X infrastructure investments made by local governments, transit agencies, real estate developers, and AV companies.

The cities that will capture the most value from this transition are those that actively manage it: that replace minimum parking requirements with maximum parking allowances, that establish dynamic curb pricing before AV PUDO demand creates chaos, that invest in V2X infrastructure ahead of AV fleet scale, and that negotiate equity service requirements into AV operating permits before the geographic inequity of owner-operator models becomes entrenched.

The cities that will lose are those that wait for AV deployment to happen to them — that watch parking garage bonds go underwater, that face AV PUDO chaos on streets designed for human drivers and static parking, and that let market forces determine which neighborhoods get AV service and which don’t.

AV urban design is not a technology story. It is a governance story. The technology is arriving regardless. What varies is whether cities shape it or are shaped by it.

Note: All figures labeled “(est.)” are derived from public disclosures, urban planning research, industry analyst estimates, and reported data as of mid-2026. This article does not constitute investment advice.

Sources

• Parking and the city — Reinventing Parking / Streetsblog ↗
• Curb management and AV fleets — NACTO ↗
• AV urban impact research — Eno Center for Transportation ↗
• V2X and smart city infrastructure — US DOT ↗
• Minimum parking reform — Parking Reform Network ↗

Tags