2026-06-18 — views
Physical AI International Expansion — Waymo's US Focus, Tesla FSD's EU Path, and China's Parallel AV Race
Waymo operates in the US only while Tesla FSD targets EU approval in 2026-2027; China runs a parallel AV race with Baidu, Huawei ADS, and WeRide already in UAE.
Article 129 in the Physical AI Benchmark Series — Physical AI International Expansion: Waymo’s US-Only Footprint, Tesla FSD’s Global Rollout Path, China’s Parallel AV Race, and Which Countries Become the Next Physical AI Markets
International expansion is the dimension that reveals the true total addressable market for Physical AI. Waymo has operated exclusively in US cities since its commercial launch. Tesla FSD is available in the US and Canada; EU approval is pending. Meanwhile, China has built a parallel autonomous-vehicle industry — Baidu Apollo, Huawei ADS, Xpeng, BYD — operating at commercial scale with different technology stacks, different regulatory frameworks, and different competitive dynamics. This article maps the global Physical AI expansion landscape as a benchmark dimension.
All figures labeled “(est.)” are derived from public market information, regulatory filings, company disclosures, and analyst estimates rather than verified primary data.
Section 1 — Waymo’s International Footprint and Expansion Signals
Waymo is the most commercially advanced AV operator in the world by driverless rides delivered, but its geographic footprint remains almost entirely domestic. Commercial service operates in Phoenix, San Francisco, Los Angeles, and Austin, with Atlanta announced. The company has not pursued international operations except for a single announced partnership with Uber in Abu Dhabi.
| Market | Status | Barrier | Timeline (est.) |
|---|---|---|---|
| United States (current) | Commercial in Phoenix, SF, LA, Austin; Atlanta announced | None — home market | Expanding now |
| Abu Dhabi, UAE | Partnership announced with Uber and Uber Eats for potential future deployment | Favorable: UAE government actively courting AV companies; no existing AV regulation to navigate | 2026-2027 pilot (est.) |
| Japan | No announcement | Regulatory: Japan requires extensive type-approval; left-hand to right-hand drive engineering change | 2028+ (est.) if pursued |
| UK / Europe | No announcement; Wayve operates in UK as a separate company | Regulatory: Euro NCAP safety standards plus EU type-approval for AV; GDPR data residency | 2029+ (est.) if pursued |
| China | Not pursuing — US-China geopolitics; Zeekr Gen 6 supplier relationship complicates data-sharing in China | Geopolitical: data sovereignty laws; Chinese AV competitors dominant | Not pursuing |
| Middle East (beyond UAE) | Saudi Arabia’s NEOM project has discussed AV integration | Favorable regulatory environment; oil wealth provides capex | 2027-2030 (est.) |
The Abu Dhabi partnership is the single international signal Waymo has sent. The UAE market is structurally attractive for a first international expansion: the government is actively creating AV-friendly frameworks, there is no entrenched legacy taxi union, and the climate (hot, dry, flat urban streets) roughly resembles Phoenix — Waymo’s best-performing market. The barrier to Japan and Europe is not just regulatory paperwork; right-hand drive vehicles require meaningful hardware re-engineering of sensor placement and actuation systems, and European GDPR requirements impose data residency obligations that would require local data infrastructure investment.
The strategic rationale for Waymo staying US-only until mid-2026 is defensible: the US market alone represents approximately 3.3 billion ride-hail trips per year (est.), and Waymo’s current 150K-plus weekly rides represent well under 1% penetration of its existing operational geographies. Domestic scale-up has higher expected value than international fragmentation for a company still expanding city by city within the United States.
Section 2 — Tesla FSD’s International Rollout Map
Tesla’s approach to FSD international expansion differs structurally from Waymo’s because FSD is software deployed onto the existing Tesla fleet — not a dedicated robotaxi fleet requiring separate vehicle deployment. Every market where Tesla sells cars is a potential FSD market, subject only to regulatory approval. This creates a much larger potential international surface area but also a more complex regulatory matrix across dozens of jurisdictions.
| Market | FSD status | Regulatory barrier | Timeline (est.) | Key signal to watch |
|---|---|---|---|---|
| United States | Full rollout; supervised; Austin robotaxi launch | Driverless permits state-by-state | Expanding now | Driverless permit expansion beyond Austin |
| Canada | Rolling out; similar regulatory framework to US | Province-by-province; generally permissive | Underway | Saskatchewan, Ontario expanding |
| European Union | FSD NOT approved — blocked by EU type-approval process (UNECE WP.29 standards) | EU requires homologation for ADAS above L2; FSD classification uncertain | 2026-2027 (est.) if approved | Euro NCAP validation; UNECE WP.29 Article 34 (L3 automated driving) |
| United Kingdom | Post-Brexit separate type-approval from EU; UK government more AV-friendly (Automated Vehicles Act 2024) | UK AV Act 2024 creates framework; Tesla could move faster in UK than EU | 2026-2027 (est.) | UK DVLA type-approval submission |
| China | Tesla manufactures in Shanghai; FSD China approval uncertain; Huawei ADS and local competitors dominate L2+ | Chinese government hesitant to approve foreign AI driving software at scale; data sovereignty | Very uncertain; 2027+ (est.) at best | MIIT approval; data localization compliance |
| Japan | Right-hand drive Tesla sold; FSD capability present but not activated | Japan’s AV regulatory framework (Road Traffic Act) requires domestic testing and certification | 2026-2027 (est.) with limited rollout | MLIT certification submission |
| South Korea | Tesla sold; FSD under regulatory review | Korean AV law relatively permissive; MOLIT approval needed | 2026-2027 (est.) | MOLIT safety certification |
| Australia | Right-hand drive Tesla sold; FSD not activated | Australian Road Rules require human responsibility; state-by-state variation | 2027+ (est.) | State-level pilot programs |
The EU market represents the single largest near-term TAM unlock for Tesla FSD. The EU has approximately 450 million residents and a vehicle fleet of roughly 250 million passenger cars (est.). UNECE WP.29 Regulation 157 governs L3 highway automation and provides a functional pathway for FSD approval — but it requires extensive homologation testing and type-approval documentation that Tesla has not yet completed in the EU system. The UK Automated Vehicles Act 2024 is globally the most comprehensive AV legislation and creates a clear liability framework (the “Authorised Self-Driving Entity” or ASDE model) that may allow Tesla to achieve UK approval faster than EU approval post-Brexit.
China is the most complex market. Tesla manufactures vehicles in Shanghai, giving it deep local supply chain and manufacturing relationships. But FSD approval in China requires compliance with Chinese data localization laws (Personal Information Protection Law plus Data Security Law), which restrict the cross-border transfer of data collected by autonomous systems — making Tesla’s cloud-based FSD improvement loop technically difficult to operate in China without dedicated local data infrastructure. Meanwhile, Huawei ADS and local competitors have already captured the L2+ ADAS market in China at scale, reducing the commercial urgency of FSD approval there.
Section 3 — China’s Parallel Physical AI Industry
China has built a parallel autonomous-vehicle ecosystem that is commercially mature in domestic markets and beginning to expand internationally. The Chinese AV industry divides broadly into two categories: robotaxi operators (Baidu Apollo Go, Pony.ai, WeRide) that follow a Waymo-like approach with lidar-heavy sensor stacks and high-definition maps; and ADAS software providers (Huawei ADS, Xpeng XNGP, BYD DiPilot) that follow a Tesla-like camera-first or camera-plus-lidar approach deployed across mass-market vehicle fleets.
| Company | Category | Current status | Key technology | International ambitions |
|---|---|---|---|---|
| Baidu Apollo Go | AV ride-hail | 6M+ driverless rides; 11+ Chinese cities; RT6 vehicle; Abu Dhabi and Hong Kong pilots announced | Lidar + camera + radar; maps-based (similar to Waymo philosophy) | Abu Dhabi pilot; Hong Kong announcement; limited by data sovereignty and geopolitics outside China |
| Huawei ADS | L2+/L3 ADAS software | Deployed in AITO, SERES, CHERY vehicles; 3M+ vehicles with Huawei ADS (est.) | Camera + radar + lidar; BEV neural net (similar to Tesla philosophy); no maps required for highways | China-domestic focus; export restricted by US sanctions on Huawei chips |
| Xpeng XNGP | L2+/L3 ADAS | Shipping in P7, G6, G9; full-city NGP (no maps) rolling out in 50+ Chinese cities | Camera + lidar; NGP mapless highway + urban | Europe expansion (Germany, Netherlands); limited by Huawei chip restrictions on export vehicles |
| BYD DiPilot | L2+/L3 ADAS | Mass market deployment across BYD fleet (3.6M vehicles sold 2025 est.); DiPilot 100/300/600 tiers | Camera + lidar (higher tiers); Momenta AI partnership | Global BYD sales in EU, AU, SE Asia mean DiPilot is internationally deployed by default |
| Pony.ai | AV ride-hail (robotaxi) | Commercial robotaxi in Beijing, Guangzhou, Shenzhen; NASDAQ IPO 2024; Saudi Arabia pilot announced | Lidar + camera + radar; maps-based | Saudi Arabia and Middle East; limited international presence outside pilots |
| WeRide | AV (robotaxi + minibus) | Commercial in Guangzhou; Abu Dhabi (UAE) commercial launch 2023 — first Chinese AV company with international commercial service | Lidar + camera + radar | UAE operational; Singapore pilot; most internationally advanced Chinese AV company |
WeRide’s Abu Dhabi commercial launch in 2023 is the most significant data point in China’s international AV expansion: it demonstrates that a Chinese AV company can operate commercially in a foreign jurisdiction ahead of Waymo’s UAE pilot. The UAE’s greenfield regulatory environment — where the government is writing the rules as it goes, rather than applying an existing framework — is the key enabler. WeRide, Baidu Apollo, and Pony.ai have all identified the Middle East as the path of least regulatory resistance for international expansion.
The Huawei constraint is significant for hardware-export scenarios. Huawei’s automotive chips (in particular the MDC driving compute platform) are subject to US export controls, which limits Huawei ADS deployment in non-Chinese markets that have US-supply-chain exposure. This creates an asymmetry: Chinese AV software can expand internationally, but Chinese AV hardware faces US export-control scrutiny.
BYD DiPilot represents the stealth international expansion vector. BYD is the world’s largest EV seller by volume, with cars sold in Europe, Southeast Asia, Australia, and Latin America. DiPilot ADAS software is bundled into BYD vehicles by default, meaning Chinese AV software is already operating on international roads wherever BYD sells cars — without the regulatory scrutiny that a dedicated robotaxi operation would face.
Section 4 — Regulatory Landscape by Region
The regulatory environment is the primary determinant of Physical AI international expansion speed. Markets with favorable regulation and greenfield frameworks (UAE, Singapore) attract early deployments regardless of market size. Markets with complex existing regulatory frameworks (EU, Japan) attract deployments only after extensive certification processes. Markets with data sovereignty requirements that conflict with cloud-based AV training loops (China for foreign companies) effectively block foreign Physical AI at scale.
| Region | AV regulatory framework | Friendliness to foreign AV | Key barrier |
|---|---|---|---|
| United States | State-by-state; CA DMV + NHTSA oversight; Texas/AZ most permissive | High for US companies (Waymo, Tesla, Aurora); mixed for foreign (Chinese AV companies face CFIUS scrutiny) | No federal framework; state fragmentation |
| European Union | UNECE WP.29 Regulation 157 (L3 highway); EU AI Act (AV = high-risk AI); type-approval homologation | Medium — detailed process but functional for US companies | GDPR data residency; UNECE WP.29 certification; EU AI Act compliance |
| United Kingdom | Automated Vehicles Act 2024 — world’s most comprehensive AV law; creates ASDE liability framework | High — AV Act signals intent; clearest liability framework globally | ASDE certification process (new, not yet tested); right-hand drive vehicle requirement |
| China | MIIT + CAAC joint AV policy; data sovereignty (PIPL + DSL); smart road infrastructure investment | Low for foreign AV companies; data sovereignty requirements block Tesla FSD approval at scale | Data localization; preferential treatment of domestic companies |
| UAE / Middle East | No existing AV law; government actively creating frameworks; ADGM as regulatory sandbox | Very high — greenfield regulatory environment; governments want AV showcase | Framework still being written; no established precedent |
| Japan | Road Traffic Act amended for AV 2023; Level 3 commercially permitted (Honda Legend first L3 car); MLIT oversight | Medium — open to AV but systematic certification required; right-hand drive | Domestic testing requirements; MLIT certification timeline |
| South Korea | Framework Act on Automated Vehicles 2020; MOLIT oversight; pilot zones | High-medium — progressive framework; active investment in AV | Domestic testing requirement; MOLIT certification |
| Singapore | LTA (Land Transport Authority) AV framework; one of most progressive globally | Very high — government-supported; small dense city ideal for AV | Small market size limits commercial appeal |
The UK Automated Vehicles Act 2024 deserves particular attention as a template for Physical AI international expansion. It establishes the Authorised Self-Driving Entity (ASDE) model, under which the company deploying the AV system — not the vehicle owner — bears legal responsibility for the vehicle’s driving. This resolves the liability ambiguity that has slowed AV deployment in many jurisdictions and provides a clear path for Tesla, Waymo, and others to seek ASDE status. If the UK ASDE certification process proves functional, the UK may be the first major European market to see commercial AV operations — ahead of the EU by 2 to 3 years (est.).
Section 5 — International Expansion Benchmark Scorecard
| Dimension | Waymo | Tesla FSD | Baidu Apollo | WeRide |
|---|---|---|---|---|
| Countries with commercial service | 1 (USA) | 2 (USA + Canada) | 1 (China) + UAE pilot | 2 (China + UAE) |
| International expansion announced | UAE (with Uber) | UK + Japan + Korea (approval pending) | Abu Dhabi, Hong Kong | UAE operational; Singapore pilot |
| China presence | None | Manufacturing only; FSD approval uncertain | Home market dominant | Home market |
| EU/UK path | Not pursuing near-term | EU 2026-2027 (est.); UK 2026-2027 (est.) | Not announced | Not announced |
| Regulatory sophistication | Proven US playbook; UAE = new | EU/UK regulatory teams engaged | China + UAE proven | UAE proven (first Chinese AV international) |
| Data sovereignty risk | Low (US HQ, US operations) | Medium (China manufacturing raises scrutiny) | High (China HQ, US-China tensions) | Medium |
| International TAM unlocked (est.) | UAE modest; global TAM remains US-first | EU 450M population = largest unlock; Japan 125M | China 1.4B = already primary TAM | UAE + SE Asia growing |
The international expansion benchmark reveals three distinct strategic postures. Waymo is deepening US scale before international; its UAE partnership is exploratory, not operational. Tesla FSD is pursuing simultaneous multi-market regulatory approval, using its existing fleet as the leverage point; EU approval would be the single largest TAM unlock available. China’s AV companies are expanding internationally via the path of least regulatory resistance — Middle East and Southeast Asia first, with WeRide holding the only confirmed international commercial operation among Chinese AV players.
The next 18 to 24 months are the decisive window for EU and UK Physical AI approval. If Tesla FSD achieves UNECE WP.29 homologation or UK ASDE certification by end of 2027, it will access a combined market of over 500 million people with high vehicle ownership rates and high willingness to pay for advanced driver assistance. If regulatory approval extends to 2029 or beyond, Chinese AV companies — particularly BYD DiPilot bundled into mass-market vehicles — will have established significant de facto Physical AI presence in European markets ahead of US competitors.
The international dimension of Physical AI is not simply a market-size calculation. It is a regulatory race, a data-sovereignty negotiation, and a geopolitical alignment question. Which countries want US Physical AI on their roads, which want Chinese Physical AI, and which want their own domestic industry — those answers will determine the global Physical AI TAM distribution for the next decade.
Note: All figures labeled “(est.)” are derived from public market information, regulatory filings, company disclosures, analyst estimates, and industry reports as of mid-2026. This article does not constitute investment advice.
Sources
- Waymo UAE partnership — Waymo blog ↗
- UK Automated Vehicles Act 2024 — UK Government ↗
- UNECE WP.29 AV regulations — UNECE ↗
- WeRide Abu Dhabi commercial launch — WeRide ↗
- Baidu Apollo international expansion — Baidu Apollo ↗