Robotaxis Hit Europe — Munich, Paris, and London Compared

For most of 2023-2025, "robotaxi" meant Waymo in Phoenix, Cruise in San Francisco, Baidu in Beijing. Europe was conspicuously absent.

That's changing fast in 2026. Three European deployments are now live with paying passengers. The contrast between them tells you a lot about how Europe is approaching autonomy differently from the US and China.

The three cities

All three are SAE Level 4, with safety driver in some operational conditions.

Munich: the cautious giant

Germany's approach is what you'd expect: engineering-first, slow, exhaustively tested.

• Three years of closed-track testing before public launch
• Mandatory teleoperation backup with sub-200ms latency requirement
• 15-page passenger consent form at first ride (digitally, but still)
• All routes pre-mapped with cm-level HD maps
• Average wait time: 8-12 minutes

The pricing is interesting: roughly 30% cheaper than a Munich taxi, with German labor laws making human taxis expensive.

Riding experience: extremely smooth, almost boringly conservative. The car will refuse certain maneuvers entirely (right-on-red equivalent, jaywalker proximity). German users seem to love it; tourists find it slow.

Paris: the political deployment

France's approach: public-private partnership with social license at the center.

• RATP (the public transit operator) owns 51% of the venture
• No layoffs allowed in the supporting bus/metro network
• Routes designed to complement, not compete with public transit
• Union approval was secured before launch (took 2 years)
• Average wait time: 5-7 minutes
• Pricing: same as Paris taxis (a deliberate policy choice)

The Parisian model is socially integrated autonomy — not "robots vs. workers" but "robots plus existing transit unions." Worth watching as a template for other unionized cities.

Operationally less sophisticated than Munich, but politically more sustainable.

London: the AI-native player

Wayve, the London-based startup, took the opposite approach from Munich's HD-map school.

• No pre-mapped routes — vehicles use end-to-end neural networks
• Trained on 6+ years of London driving from camera-equipped fleets
• Faster updates — model improvements roll out weekly
• Wider geofence: most of zones 1-3
• Average wait time: 4-6 minutes
• Pricing: 20% premium over black cabs (UK regulator allows this)

This is the most architecturally ambitious of the three. If Wayve's approach works at scale, it dramatically reduces the cost of expanding to new cities (no HD mapping needed). If it doesn't, edge cases will pile up.

Six months in, the early data suggests it's working — but London is also one of the most CCTV-dense cities in the world, which probably helps training.

Regulatory contrasts

Notable: Europe's GDPR creates ride-data challenges that US operators don't face. Every robotaxi captures passenger video by default; the consent and retention rules add operational complexity.

What surprised European riders

Based on early user surveys (n=8,400 across three cities):

• Quietness was the #1 mentioned positive — no driver chatter, no radio
• #1 complaint: cars taking weird detours to avoid construction zones
• Surprisingly low novelty curve — after 3 rides, people stop filming
• Strong preference for pre-booking vs. hail-on-demand (35% vs. 12% in US)
• 80% of riders booked solo — autonomy correlates with private rides

The "people will share cars with strangers" hypothesis is failing in Europe just like the US.

What's coming next

Announced deployments for late 2026 / early 2027:

• Amsterdam (Mobileye, similar to Munich model)
• Stockholm (Volvo + Aurora partnership)
• Madrid (Cruise return, after restructuring)
• Hamburg (Mobileye, freight-focused not passenger)
• Geneva (luxury small-fleet, Mercedes EQS robotaxi)

Notable absences: Italy, Spain, Eastern Europe. Regulatory capacity gap, not technical.

The 5-year European trajectory

Three predictions for 2030:

1. Robotaxi rides will be ~15% of urban for-hire trips in capital cities
2. The Munich model (HD maps) will lose to the London model (neural networks) once scale hits
3. A unified EU regulatory framework will replace the city-by-city patchwork around 2028

Europe started late, but the regulatory infrastructure they're building may turn out to be more durable than the US patchwork. The slow start pays off when scale arrives.

For tech tourists in Europe in 2026, riding a robotaxi in all three cities is now a doable weekend tour. Recommended.

Related: eSIM for the Europe robotaxi tour

The booking apps and payment flows all need data. Pre-load an eSIM that works across EU/UK before you fly — much cheaper than roaming and no SIM swap at the airport.