2026-06-18 👁️ — views

Physical AI Partnerships & Ecosystem Index — Who Builds the Fleet That Scales

Manufacturing partners, fleet ops, and distribution deals that determine how fast Waymo and Tesla can actually put autonomous vehicles on the road.

Article 16 in the Physical AI Benchmark Series

Articles 1 through 15 mapped the technology stack, regulatory gates, capital structure, supply-side constraints, and consumer demand behind the autonomous vehicle ramp. This article asks the question that underlies all of them: who actually manufactures the vehicles, manages the fleets, and distributes the rides? The supply chain behind physical AI is the hidden variable in every scale forecast.

Section 1 — Manufacturing Partnership Table

The vehicle itself is the foundational unit of AV fleet economics. Before a robotaxi can earn revenue, it has to be built, delivered, mapped, and onboarded. Manufacturing capacity sets the ceiling on how fast any operator can grow.

All capacity figures are estimates based on public announcements and industry reporting. Neither Waymo nor Zeekr has disclosed exact contracted volumes.

Section 2 — Fleet Operations Partners

Building the vehicle is necessary but not sufficient. Fleet management, servicing, and charging infrastructure determine the operating cost per mile once vehicles are deployed.

The Moove partnership is structurally important for Waymo’s expansion economics. By outsourcing fleet management to a specialist operator, Waymo can enter new cities without building a full service operation from scratch. The cost is margin share; the benefit is speed and capital efficiency.

Tesla’s owner-enrollment model inverts this logic entirely. Rather than managing a centrally-owned fleet, Tesla collects a platform fee on rides generated by owner-deployed vehicles. Owners absorb the depreciation and maintenance. Tesla scales without deploying capital.

Section 3 — The Zeekr-Waymo Manufacturing Dependency

Waymo’s Gen 6 vehicle is manufactured by Zeekr, a premium EV brand under Geely Group based in China. This is the most consequential single partnership in the Waymo supply chain.

Why Zeekr? Zeekr brings high-volume EV manufacturing expertise, purpose-built vehicle architecture for AV sensor integration, and manufacturing costs that would be substantially higher if replicated in the United States. Waymo gets a production-grade vehicle without building or owning a factory.

The geopolitical risk: US-China trade tensions create a concentration risk at the foundation of Waymo’s fleet scaling plan. Potential disruption vectors include:

• Tariffs on AV components or vehicles manufactured in China
• Export controls on lidar, radar, or compute components flowing between the two countries
• Diplomatic deterioration that complicates contractual manufacturing relationships

Waymo has not disclosed any backup manufacturing partner or domestic production contingency. If the Zeekr supply line is disrupted — by tariff escalation, regulatory action, or geopolitical event — Waymo’s Gen 6 production timeline would slip with no immediate fallback. This is a single-point-of-failure risk that does not appear on most AV analyst scorecards.

Estimated timeline: Gen 6 production began in 2025. Ramp to 10,000+ units per year is the key milestone that would allow Waymo to meaningfully expand beyond its current four operational markets. At 20,000 units per year — the high end of estimates — Waymo could deploy roughly 4,000–5,000 net new vehicles annually after accounting for attrition in the existing fleet.

Section 4 — Tesla’s Vertical Integration Advantage

Tesla manufactures its own vehicles, batteries, chips (HW4), and AI training infrastructure (Dojo) — all in-house. No manufacturing partner. No component sourcing risk from a geopolitically exposed supplier.

This vertical structure changes the risk profile of Tesla’s AV scaling plan in three ways:

1. No partner dependency. Waymo’s fleet ceiling is set by Zeekr’s production allocation and willingness to expand capacity. Tesla’s ceiling is set by its own Gigafactory throughput — which it controls and has invested tens of billions of dollars to expand.

2. Dual-vehicle strategy. The Cybercab is a purpose-built robotaxi targeting $25K–$30K unit cost. The Model Y robotaxi requires no new manufacturing at all — it is a software activation on top of an existing production vehicle. This means Tesla can begin generating robotaxi revenue before the Cybercab factory line is even running, using current production capacity.

3. Charging infrastructure is pre-built. With 60,000+ Superchargers globally, Tesla already has the charging backbone for a large-scale AV fleet. Waymo must negotiate third-party charging contracts or rely on Moove for each new market. This is not a theoretical advantage — charging reliability is a direct driver of fleet uptime and per-mile economics.

The primary constraint on Tesla’s AV scaling is not manufacturing capacity. It is the FMVSS regulatory waiver required to operate the Cybercab (which lacks a steering wheel and pedals) at commercial scale. Article 13 in this series covered that gate in detail. Until the waiver is obtained, Cybercab production volumes are moot.

Section 5 — Partnership Ecosystem Scorecard

Cost per vehicle figures are estimates. Waymo Gen 6 cost has not been disclosed; the range reflects analyst estimates based on comparable AV-grade sensor and compute bills of materials. Tesla Cybercab cost reflects stated targets, not confirmed production economics.

Reading the scorecard: Tesla’s vertical integration advantage is structural and durable. It compounds over time — every dollar invested in Gigafactory capacity, Supercharger infrastructure, and HW4 chip development is a fixed cost that spreads across an ever-larger fleet. Waymo’s outsourced model is more capital-light in the short term but carries concentration risk (Zeekr) and per-vehicle cost disadvantage at scale.

Waymo’s counter-advantage — broader distribution today, via Uber and Google Maps — is real but temporary. If Tesla achieves its stated ride cost targets, distribution reach becomes less important than price and availability.

How This Article Fits the Series

This is article 16 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: The demand side — rider experience, adoption curves, and pricing
• Article 16 (this article): The supply chain — manufacturing partners, fleet operations, and distribution ecosystem

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

Sources

• Waymo Gen 6 vehicle and Zeekr partnership — Waymo blog ↗
• Waymo Moove fleet operations partnership — Waymo press ↗
• Waymo x Uber distribution deal — Waymo press ↗
• Tesla Cybercab manufacturing plans — Tesla AI ↗
• Tesla Supercharger network — Tesla ↗

Tags