2026-06-18 — views

Physical AI International Expansion 2026 — Waymo US-Only vs Tesla Global FSD Fleet: Europe, China, and the Global Ramp Benchmark

Waymo is US-only across 4 cities. Tesla has FSD-capable vehicles in 50-plus countries — a structural global Physical AI edge Waymo cannot match.

Article 175 in the Physical AI Benchmark Series — International Expansion and the Global Ramp Index

The Physical AI benchmark series has examined Waymo and Tesla across ramp speed, fleet depth, unit economics, software architecture, regulatory pathways, and capital structure. One dimension has received less attention: international reach. As of mid-2026, Waymo has zero commercial operations outside the United States and no announced near-term international launch. Tesla sells FSD-capable vehicles in more than 50 countries, has an existing installed base in Europe and China, and is pursuing regulatory approval to unlock driverless capability on that global fleet. This article quantifies the international dimension of the Physical AI race.

Section 1 — Waymo’s International Position: US-Only, No Near-Term International Ramp

Waymo operates commercial driverless paid rides in four US cities. Its geographic footprint is entirely domestic.

Dimension	Status Q2 2026	Notes
International commercial operations	Zero — all 4 commercial driverless markets are US cities (Phoenix, SF, LA, Austin TX)	Waymo has no announced international commercial service
EU operations	No EU commercial service; no EU testing permit publicly announced	Waymo has focused 100% on US markets; EU regulatory complexity (type approval, GDPR, varying national laws) makes EU entry challenging
UK operations	No UK commercial service	Post-Brexit UK has its own AV regulatory framework (AEVA 2018, AV Act 2024)
China operations	No China operations; geopolitical complexity makes China entry for a Google/Alphabet subsidiary essentially impossible	A Google-affiliated company operating autonomous vehicles and collecting mapping data in China would face severe regulatory and national security obstacles
Japan/Korea/Australia	No operations	No announced plans
Why Waymo stays US-focused	Each new geography requires: new HD map generation, new regulatory approval, new ROC infrastructure, new fleet logistics; Waymo’s capital is better deployed deepening US markets where it already has permits	Waymo’s international expansion constraint is operational (each city requires significant investment) and regulatory (each country is a new process)
Competitive threat from international AV companies	Baidu Apollo (China), Pony.ai (China, US-listed), WeRide (China, US-listed), Zoox (Amazon, US), Nuro (US, delivery), Mobileye (Intel spinoff, Jerusalem pilot) are all expanding; none has Waymo’s US commercial depth	Waymo’s US market position is defensive: competitors must enter Waymo’s core markets to challenge it; Waymo doesn’t need to enter their markets to defend
International timeline (est.)	No near-term international commercial launch announced; if Waymo eventually expands internationally, UK or Australia most likely first (English-speaking, right-hand or left-hand drive compatible, permissive AV frameworks) (est.)	UK AV Act 2024 created a clearer framework for AV commercial operation than exists in most EU member states; could be Waymo’s first international market (est.)

The structural constraint. Waymo’s business model requires building a complete physical and digital stack in every geography it enters: HD maps generated city-block by city-block, a teleoperator center staffed and connected, a fleet servicing depot with charging and calibration infrastructure, and a country-specific regulatory approval. Each of these steps takes 12 to 24 months minimum per city. Doing this in a foreign country — with a foreign language, foreign regulatory agencies, foreign insurance and liability law, and foreign HD mapping requirements — multiplies the cost and timeline by a material factor. Waymo has chosen, rationally, to redeploy that capital into deepening its US commercial position rather than attempting international geographic expansion in a phase where the US alone still represents a multi-decade growth runway.

Section 2 — Tesla FSD International: Europe Supervised, China Strategic

Tesla’s international position is structurally different from Waymo’s because Tesla sells hardware (vehicles) globally before software capability (driverless) is approved. Every FSD-capable Tesla sold internationally is a potential revenue unit the moment regulatory approval is granted — without the physical fleet deployment step Waymo requires.

Market	FSD status Q2 2026	Regulatory framework	Revenue potential	Key challenge
Europe (EU/EEA)	FSD supervised available in select EU markets (Germany, France, Norway est.); full EU rollout pending type approval and regulatory approvals in each member state	EU type approval under UNECE WP.29; automated driving functions covered under UN Regulation No. 157 for ALKS (Automated Lane Keeping Systems); full FSD beyond ALKS requires additional approvals	Large TAM: est. 250M-plus vehicles on EU roads; even low FSD penetration = significant revenue (est.)	EU regulatory process is slower and more complex than US; requires engagement with EU Commission and national type approval authorities; GDPR adds compliance layer
Germany (specific)	Germany’s AFGBV (Autonomous Driving Act, 2021) allows Level 4 AV operation in defined areas; first EU country to create a specific Level 4 legal framework	AFGBV: requires Federal Motor Transport Authority (KBA) approval for specific operational design domains (ODDs); not a blanket approval for all roads	Germany is the most permissive EU country for AV; BMW, Mercedes, and VW have filed for AFGBV approvals; Tesla could potentially use AFGBV framework for FSD limited deployment (est.)	AFGBV requires KBA approval per ODD; limited geographic scope; not a path to city-wide FSD as in US
China	Tesla has been seeking approval to offer FSD in China; requires Chinese government data localization approval (all data collected in China must stay in China); discussions with regulators ongoing as of mid-2026	China’s Smart Connected Vehicle (SCV) regulations require strict data localization; MIIT oversees AV approval; vehicles must map and test in China with Chinese regulatory oversight	China is Tesla’s 2nd largest market; FSD attach rate in China could be significant if approved; Chinese consumers are receptive to software features	Data localization: Tesla’s FSD training pipeline relies on sending video to US servers; China requires local data storage; Tesla is negotiating a China-compliant data architecture; not yet approved as of mid-2026
UK	FSD supervised available; UK-specific type approval post-Brexit; UK AV Act 2024 creates framework for automated driving	UK AV Act 2024: establishes legal liability for “Automated Driving Systems” and creates a regulatory regime for self-driving vehicles; more flexible than EU type approval	UK is Tesla’s largest European market by vehicles; FSD already available supervised; driverless path clearer than EU under AV Act 2024	UK regulatory approval for driverless FSD still required under AV Act 2024; not yet granted
Japan	Tesla sells vehicles in Japan; FSD supervised limited; Japanese regulatory framework for AV (Road Traffic Act amendments)	Japan amended Road Traffic Act (2022) to allow Level 4 AV in limited areas; MLIT oversight	Japan has high Tesla adoption in EV segment; FSD revenue potential moderate	Japanese roads are complex: narrow streets, unique signage, left-hand traffic; FSD’s training data is predominantly US/EU right-hand traffic; additional Japan-specific training required
Australia	Tesla sells vehicles in Australia; FSD supervised limited; Australian framework less developed	Australia’s National Transport Commission (NTC) oversees AV policy; state-by-state approach	Australian roads similar to UK (left-hand drive); potential for early FSD driverless approval if UK framework serves as model	Less market urgency than EU/China; moderate revenue potential

The international optionality argument. Every FSD-capable Tesla delivered internationally is a future revenue option. When regulatory approval arrives — for Germany’s AFGBV-governed geofences, for China’s data-localized zones, for the UK’s AV Act framework — Tesla does not need to manufacture and deploy new vehicles. The fleet is already there, already owned by customers, and can receive driverless capability via over-the-air software update. Waymo’s equivalent event would require deploying physical fleet from zero in that country. Tesla’s international optionality is embedded in every FSD vehicle already sold globally.

Section 3 — EU AV Regulatory Framework: Why It Differs from the US

The EU regulatory environment for autonomous vehicles is not simply “US rules translated to European languages.” It is a structurally different system with different jurisdictional architecture, different data privacy requirements, and a different liability framework.

Regulatory dimension	US approach	EU approach	Impact on Tesla/Waymo
Primary framework	State-by-state AV laws plus NHTSA FMVSS (federal safety standards); no unified federal AV approval for Level 4	UNECE WP.29 Automated Driving regulations plus EU type approval process; applies uniformly across EU member states once approved	EU type approval is a single process covering all 27 member states — more work upfront but broader coverage once achieved
Level 4 AV approval	No standardized Level 4 approval; NHTSA FMVSS exemption process covers vehicle safety standards only; states handle operational permits	UNECE WP.29 does not yet have a complete Level 4 framework; Germany’s AFGBV is the only EU-level Level 4 operational framework (limited ODDs)	Level 4 driverless in EU is effectively limited to Germany’s AFGBV framework currently; no EU-wide Level 4 AV path exists yet
Data privacy	CCPA (California) plus limited federal data privacy; AV data collection regulated at state level	GDPR applies to all AV data collected in EU; right of explanation, data minimization, purpose limitation all apply to AV sensor data	Tesla’s Data Engine (sending video from EU vehicles to US for training) faces GDPR scrutiny; requires either EU-local data processing or explicit consent frameworks
Type approval process	Vehicle safety: FMVSS compliance (NHTSA); no separate AV type approval	Vehicle safety: EU type approval covers vehicle plus AV system as integrated unit; KBA (Germany), DREAL (France), DVSA (UK post-Brexit) are national approval authorities	EU type approval for AV functions takes longer than FMVSS compliance; each new FSD version requires regulatory re-assessment
Liability framework	US tort system; liability varies by state; no unified AV liability law	EU Product Liability Directive (2024 revision) explicitly covers AI/AV systems; clearer liability attribution	EU’s explicit AI/AV liability framework may actually improve commercial deployment confidence once approved

GDPR as a data flywheel constraint. Tesla’s competitive moat in FSD development is its fleet learning loop: hundreds of thousands of vehicles send video clips of edge cases back to Tesla’s training servers, which trains the next FSD version, which is pushed back to all vehicles via OTA. GDPR restricts this in the EU. Tesla must either obtain explicit GDPR-compliant consent from EU vehicle owners for video telemetry collection, build EU-local data processing infrastructure, or operate with a reduced EU training data contribution. This is not a fatal constraint — Tesla has significant non-EU training data — but it means the EU fleet contributes less to the global FSD improvement loop than the US fleet does per vehicle.

Section 4 — China AV Market: Domestic Competition and Tesla’s Data Dilemma

China presents a unique combination: the world’s largest EV market, a domestic AV ecosystem more mature than Europe’s, and a regulatory environment that creates specific structural challenges for foreign AV companies.

Dimension	Detail	Notes
China domestic AV leaders	Baidu Apollo: est. 700-plus driverless robotaxi vehicles in China (est.); operating commercially in Beijing, Chongqing, Wuhan, Shenzhen; Pony.ai (NYSE: PTC): est. 250-plus vehicles, commercial robotaxi in Beijing, Guangzhou; WeRide: operations in 7-plus Chinese cities	China’s domestic AV companies have more deployed vehicles than Waymo in China; in the Chinese market, Waymo is the laggard by design (not operating at all)
Baidu Apollo commercial scale (est.)	est. 700-plus driverless commercial robotaxi vehicles (est.); est. millions of driverless miles in China (est.)	Baidu Apollo is arguably the closest Chinese equivalent to Waymo; operates commercially in defined zones across major Chinese cities
Tesla’s China FSD challenge	Tesla’s FSD training relies on video telemetry sent to US servers; China’s PIPL (Personal Information Protection Law) and MIIT regulations require AV data to stay in China; Tesla must build China-local data infrastructure to comply	China data localization is the fundamental blocker for FSD in China; Tesla has been in negotiations but resolution not confirmed as of mid-2026
Tesla China market importance	China is est. 20–25% of Tesla’s total vehicle revenue (est.); FSD attach revenue in China at even 10% attach rate = significant; Chinese consumers historically receptive to software subscriptions	The FSD China opportunity is large but data localization creates a compliance architecture challenge Tesla must solve before launch
Geopolitical risk for both companies	Waymo (Alphabet/Google) cannot operate in China due to Google’s restricted status; Tesla operates in China but faces periodic US-China trade tension headwinds	Waymo’s China constraint is structural and permanent (Google is blocked); Tesla’s is regulatory/compliance and potentially solvable
What China data localization means for FSD training	If Tesla isolates China data in China, it cannot use it to improve the global FSD model; OR Tesla builds a separate China-local training pipeline, which duplicates infrastructure cost	The data architecture decision for China FSD is a significant infrastructure investment that Tesla has not publicly confirmed undertaking as of mid-2026

The Waymo China impossibility. Waymo is a subsidiary of Alphabet, which owns Google. Google’s services are blocked in China. A Google-affiliated company attempting to operate autonomous vehicles on Chinese roads — vehicles that generate HD maps, collect continuous sensor data of streets and buildings, and transmit that data to servers — would face insurmountable national security objections from Chinese regulators. This is not a solvable regulatory challenge; it is a structural geopolitical barrier. Waymo’s China market is permanently zero.

Section 5 — Global Physical AI Ramp Index: International Dimension

KPI	Waymo global	Tesla global	H2 2026 catalyst	2028 est.
Countries with commercial driverless service	1 (USA)	0 driverless (supervised FSD in 10-plus countries)	Waymo: UK as possible 1st international (est. low probability H2 2026); Tesla: FSD China approval possible H2 2026/H1 2027	Waymo est. 1–2 countries (est.); Tesla: FSD driverless in UK/Germany possible (est.)
International FSD-capable vehicles on road	0	est. 500,000-plus FSD-capable vehicles in EU alone (est.); est. 300,000-plus in China (est.)	Tesla: any new FSD driverless approval unlocks massive existing fleet	Tesla: est. 2–3M international FSD-capable fleet by 2028 (est.)
China commercial status	Not possible (Google restricted)	FSD supervised approval pending; driverless est. 2027-plus	Tesla FSD China approval = major revenue catalyst	Tesla: China FSD est. $1B-plus annual revenue potential at scale (est.)
EU driverless status	Zero EU operations	FSD supervised; driverless via Germany AFGBV or EU type approval	AFGBV limited ODD approval for Tesla (possible H2 2026 est.)	EU Level 4 framework still developing; full driverless EU-wide est. 2028–2030
International competitive threat to Waymo	Chinese AV companies (Baidu Apollo, Pony.ai) expanding internationally; potential EU pilots	Chinese AV companies in China are direct Tesla FSD competitors in Chinese market	No imminent threat to Waymo’s US operations from Chinese AV companies	Baidu Apollo international expansion could accelerate; watch for EU pilots

Ramp index verdict — international dimension. Tesla holds a structural international advantage Waymo cannot match through fleet deployment alone: Tesla has sold FSD-capable vehicles in 50-plus countries, building a global installed base that can receive driverless capability via OTA when approved. Waymo must physically deploy fleet, ROC infrastructure, and HD maps in each new country — a capital-intensive, country-by-country process that starts from zero in every new geography. Tesla’s international optionality is embedded in every FSD vehicle already sold globally. If Tesla achieves driverless approval in Germany’s AFGBV framework and China’s data-localized zones by 2027, the revenue unlock comes from existing fleet — no new vehicles required.

The one dimension where this advantage narrows: Waymo’s driverless technology is already proven at commercial scale. Tesla’s FSD supervised is not yet driverless in any market. The international vehicle install base advantage is only monetizable when Tesla achieves driverless regulatory approval — which it has not yet achieved anywhere as of mid-2026.

Section 6 — Key Metrics to Watch

Tesla FSD China regulatory update. Any public statement from MIIT or Tesla confirming data localization architecture approval would be the single highest-signal event for Tesla’s international FSD revenue timeline. Watch for: Shanghai data center expansion announcements, China-specific FSD version rollout, or Tesla China press release on regulatory status.

Germany AFGBV application by Tesla. A Tesla filing with KBA (Kraftfahrt-Bundesamt) for an AFGBV operational design domain approval would signal that Tesla is actively pursuing the EU’s most accessible Level 4 path. No such filing has been publicly confirmed as of mid-2026.

UK AV Act regulatory timeline. The UK’s Centre for Connected and Autonomous Vehicles (CCAV) is expected to publish implementation guidance under the AV Act 2024. Any announcement of an approval pathway for Level 4 self-driving vehicles in the UK would be positive for Tesla (large UK installed base) and potentially for Waymo (most accessible first international market).

Baidu Apollo international expansion. Baidu has disclosed interest in international expansion, particularly in Middle East and Southeast Asia. Any commercial launch outside China would be the first head-to-head test between Chinese and US AV companies in a neutral market.

Section 7 — About This Series

This is article 175 in the Physical AI Benchmark Series. The series has covered the ramp index, humanoid robot race, regulation, capital, compute, sensors, unit economics, fleet operations, HD mapping, software architecture, safety data, investor frameworks, city-by-city expansion, Cybercab versus Model Y unit economics, Optimus manufacturing, valuation and IPO analysis, the Waymo-Uber partnership, Gen 6 vehicle transition, and the full 2028 Bear/Base/Bull forecast — among many other dimensions.

This article adds the international dimension: the geographic reach of the two leading Physical AI companies as of mid-2026, the regulatory frameworks governing international expansion, and the structural asymmetry between Tesla’s global fleet optionality and Waymo’s city-by-city physical deployment requirement.

The central finding: international expansion is the single dimension of the Physical AI race where Tesla holds the clearest structural advantage over Waymo — an advantage rooted not in software superiority but in the simple fact that Tesla sells hardware globally while Waymo deploys service geographically. Resolving this asymmetry requires Waymo to make a multi-billion-dollar bet on international physical infrastructure. There is no indication Waymo is making that bet in the near term.