AVFox
Loading AVFox content.
← All rankings
Autonomous mobility
Demand
Which apps book the most autonomous rides, whoever built the driving system or owns the cars.
How to read the scores. Every AVFox ranking uses the same yardstick: 10 means 10 billion autonomous trips or deliveries a year, about the number of rides Uber completes worldwide, and 3 marks the start of driverless commercial service. Each point in between is meant to take about the same effort to climb. The score measures one thing: how many autonomous trips or deliveries a company completes today, not its potential, profitability, safety or quality of service. Few companies publish these numbers, so most scores are estimates built from public information.
- Geography
- Global, shown excluding and including China
- Status
- Research preview
- Methodology
- Demand methodology v0.1
Research preview
Research preview under methodology v0.1, using public information available through November 8, 2026. Few platforms report autonomous rides by app, so most figures here are our estimates from fleets, utilization and market share. The information is imperfect, so the ranking is too: read the scores as an informed approximation, not an exact count. All scores are estimates.
Research cutoff: November 8, 2026
Excluding China
| Rank | App | Deployment stage | Score (out of 10) | Summary |
|---|---|---|---|---|
| 1 |  WaymoDriverless commercial Waymo's own app books most of its driverless rides; Uber and Lyft book them in three cities. | Driverless commercial | 5.0 | Waymo's own app books most of its driverless rides; Uber and Lyft book them in three cities. |
How we got hereHow the score was calculatedWaymo's total from the ADS ranking is 30.3M trips a year (estimated). About 400 cars in Austin and Atlanta are booked only through Uber; at Uber's reported 30% utilization edge, that's 3.8M trips. Nashville's fleet is shared with Lyft, half each: 0.2M for Lyft. Waymo app: 30.3M − 3.8M − 0.2M = 26.3M (estimated) → 5.02, shown as 5.0. FootprintDriverless public rides in the Waymo app in its US service cities. Austin and Atlanta are booked only through Uber; Nashville through the Waymo app and Lyft. What would move the scoreA new ride figure from Waymo, or one that separates rides by app. 5.1 needs about 27.7M trips a year booked in the Waymo app. | ||||
| 2 |  UberDriverless commercial The biggest third-party app for robotaxi rides: Waymo in Austin and Atlanta, WeRide and Apollo Go in the UAE. Uber also offers rides with onboard safety operators in multiple markets and has announced further partnerships. | Driverless commercial | 4.3 | The biggest third-party app for robotaxi rides: Waymo in Austin and Atlanta, WeRide and Apollo Go in the UAE. Uber also offers rides with onboard safety operators in multiple markets and has announced further partnerships. |
How we got hereHow the score was calculatedWaymo: about 400 Uber-only cars in Austin and Atlanta × 1.3 (Uber's utilization edge) × 7,362 trips per car a year = 3.8M. UAE: about 150 driverless cars × 17.5 trips a day = 1.0M. Total 4.8M (estimated) → 4.28, shown as 4.3. Rides from supervised partners (Avride, Verne, Riyadh) don't count. FootprintDriverless rides in Austin, Atlanta, Abu Dhabi and Dubai; supervised partner rides in other cities. Zoox and Motional launches are due later in November. What would move the scoreUber's UAE driverless count: at the nearly 200 it expects by year-end, its score would rise from 4.28 to 4.31, still shown as 4.3. | ||||
| 3 |  Tesla RobotaxiDriverless commercial Tesla's own app, the only way to book its driverless rides. | Driverless commercial | 3.8 | Tesla's own app, the only way to book its driverless rides. |
How we got hereHow the score was calculatedTesla reports miles, not trips. About 1.65M driverless miles in Q3 2026 × 4 ÷ 6.7 miles per trip = about 1.0M trips a year (estimated) → 3.76, shown as 3.8. The Tesla Robotaxi app is the only way to book them. Footprint6 driverless public cities; supervised service in the Bay Area; 4 cities announced. 300+ driverless vehicles. What would move the scoreReported trip counts. If October's pace holds, the score would be about 4.1. | ||||
| 4 |  ZooxDriverless commercial Zoox's own app; the only way to book its robotaxis until the Uber launch. | Driverless commercial | 3.6 | Zoox's own app; the only way to book its robotaxis until the Uber launch. |
How we got hereHow the score was calculatedZoox reports riders, not trips. About 100 vehicles × 5,000 trips a year = 0.5M (estimated) → 3.59, shown as 3.6. All are booked in the Zoox app until its Uber launch. Footprint2 driverless public cities; about 10 testing cities; Los Angeles announced for 2027. About 100 vehicles (press reports). What would move the scoreA reported fleet or trip count. At its March–August rider pace (about 1.2M trips a year) it would score 3.8. Rides booked through Uber from late November will count for Uber. | ||||
| 5 |  AutoGoDriverless commercial Abu Dhabi app that books Apollo Go's driverless robotaxis. | Driverless commercial | 3.4 | Abu Dhabi app that books Apollo Go's driverless robotaxis. |
How we got hereHow the score was calculatedApollo Go's Abu Dhabi rides are booked through AutoGo. About 50 cars (estimated: 'dozens' at launch, 'hundreds' targeted for 2026) × 5,000 trips a year = 0.25M → 3.41, shown as 3.4. Tied with Apollo Go's Dubai service; AutoGo is listed first because its service started earlier (January vs March 2026). FootprintAbu Dhabi, starting on Yas Island; driverless public rides since January 2026. What would move the scoreA published fleet count. At 150 cars it would score 3.7. | ||||
| 6 |  Apollo GoDriverless commercial Baidu's robotaxi app. This list counts only its trips in Dubai, its one service outside China. | Driverless commercial | 3.4 | Baidu's robotaxi app. This list counts only its trips in Dubai, its one service outside China. |
How we got hereHow the score was calculatedCounts only rides outside mainland China. Apollo Go's own app books driverless rides in Dubai, launched with Dubai Taxi Company on March 31, 2026 with about 50 cars: 50 × 5,000 trips a year = 0.25M (estimated) → 3.41, shown as 3.4. Its Abu Dhabi rides are booked through AutoGo, and Dubai rides booked through Uber count for Uber. Tied with AutoGo, which is listed first because its service started earlier (January vs March 2026). FootprintDubai, with Dubai Taxi Company, since March 2026. Its rides in China count only in the Including China list. What would move the scoreA published Dubai fleet or ride count. Dubai plans to grow the service toward 1,000 cars over the coming years. | ||||
| 7 |  LyftDriverless commercial Books Waymo rides in Nashville from a fleet shared with the Waymo app. | Driverless commercial | 3.3 | Books Waymo rides in Nashville from a fleet shared with the Waymo app. |
How we got hereHow the score was calculatedWaymo's Nashville fleet, 'dozens of vehicles' taken as 50, is shared by the Waymo app and Lyft, half each: 25 cars × 7,362 trips a year = 0.18M (estimated) → 3.33, shown as 3.3. May Mobility rides in Atlanta are supervised and don't count. FootprintNashville (Waymo, since September 9, 2026); supervised May Mobility rides in Atlanta. What would move the scoreA published Nashville fleet size. Lyft expects it to grow significantly by year-end; at 100 cars it would score 3.5. | ||||
| 8 |  MOIASupervised rides Volkswagen's ride-pooling app; autonomous rides in Hamburg with a safety driver. | Supervised rides | 2.2 | Volkswagen's ride-pooling app; autonomous rides in Hamburg with a safety driver. |
How we got hereHow the score was calculated1.5 for autonomous rides in the app + 0.4 for public riders (selected residents) + 0.2 for an area, not a fixed route + 0.1 for a dated driverless target (2027) = 2.2. Which app riders use isn't stated; we assume MOIA's. FootprintSelected areas of Hamburg, up to 10 vehicles, since July 2026. What would move the scoreRemoving the safety driver (+0.4), service all week for 12+ hours a day (+0.2), or 25+ vehicles (+0.1). | ||||
| 9 |  VerneSupervised rides Croatian robotaxi app; paid rides in Zagreb with a safety operator, also bookable on Uber. | Supervised rides | 2.2 | Croatian robotaxi app; paid rides in Zagreb with a safety operator, also bookable on Uber. |
How we got hereHow the score was calculated1.5 for paid rides in the app + 0.4 for public riders + 0.2 for an area (city centre and airport) + 0.1 for a driverless target by end of 2026 = 2.2. Tied with MOIA, which is listed first on vehicles (10 vs not disclosed). FootprintZagreb city centre and airport, since April 2026. What would move the scoreRemoving the safety operator (+0.4), confirmed service of 12+ hours a day (+0.2), or 25+ vehicles (+0.1). | ||||
| 10 |  Grab (Ai.R)Supervised rides Grab's autonomous service in Punggol, Singapore, booked through its Ai.R portal rather than the Grab app. | Supervised rides | 1.9 | Grab's autonomous service in Punggol, Singapore, booked through its Ai.R portal rather than the Grab app. |
How we got hereHow the score was calculated1.5 for autonomous rides + 0.4 for public riders = 1.9. Three fixed routes, weekdays 9:30–17:30, 10 vehicles, with a safety operator aboard. FootprintPunggol, Singapore. What would move the scoreRemoving the safety operator (+0.4), service all week for 12+ hours a day (+0.2), an area instead of fixed routes (+0.2), or 25+ vehicles (+0.1; 50 are planned by March 2027). | ||||
| 11 |  BoltAnnounced Testing Pony.ai robotaxis with Stellantis in Luxembourg; no rider bookings yet. Bolt also plans at least 25,000 Lucid robotaxis using NVIDIA technology in Europe; these are announced, not deployed. | Announced | 0.8 | Testing Pony.ai robotaxis with Stellantis in Luxembourg; no rider bookings yet. Bolt also plans at least 25,000 Lucid robotaxis using NVIDIA technology in Europe; these are announced, not deployed. |
How we got hereHow the score was calculated0.4 for a signed agreement naming a market (Pony.ai and Stellantis, Luxembourg) + 0.4 for vehicles deployed for the pilot = 0.8. No permit for passenger service and no dated rider launch. FootprintLuxembourg (testing). What would move the scoreA passenger-service permit (+0.3) or a dated rider launch (+0.3). Rides in the app move it to 1.5 or higher. | ||||
| 12 |  inDriveAnnounced Agreements to bring robotaxis to Riyadh and Astana; no fleet or launch date yet. | Announced | 0.4 | Agreements to bring robotaxis to Riyadh and Astana; no fleet or launch date yet. |
How we got hereHow the score was calculated0.4 for agreements naming Riyadh (December 2025, with WeRide's Saudi dealer AI Driver) and Astana (October 2026, with Pony.ai and NVIDIA). No committed fleet, permit or dated rider launch. FootprintRiyadh and Astana (planned). What would move the scoreA committed fleet (+0.4), a permit (+0.3) or a dated launch (+0.3). | ||||
Autonomous rides only or mainly in the EU
Volume unknown
Driverless rides in the app, but not enough public information to estimate trips. Not scored.
Via
Waymo vehicles in Via's on-demand transit service in Chandler, Arizona. Volume isn't disclosed.
Including China
Every autonomous ride, including apps that operate only in China.
| Rank | App | Deployment stage | Score (out of 10) | Summary |
|---|---|---|---|---|
| 1 | Waymo Driverless commercial Waymo's own app books most of its driverless rides; Uber and Lyft book them in three cities. | Driverless commercial | 5.0 | Waymo's own app books most of its driverless rides; Uber and Lyft book them in three cities. |
How we got hereHow the score was calculatedWaymo's total from the ADS ranking is 30.3M trips a year (estimated). About 400 cars in Austin and Atlanta are booked only through Uber; at Uber's reported 30% utilization edge, that's 3.8M trips. Nashville's fleet is shared with Lyft, half each: 0.2M for Lyft. Waymo app: 30.3M − 3.8M − 0.2M = 26.3M (estimated) → 5.02, shown as 5.0. FootprintDriverless public rides in the Waymo app in its US service cities. Austin and Atlanta are booked only through Uber; Nashville through the Waymo app and Lyft. What would move the scoreA new ride figure from Waymo, or one that separates rides by app. 5.1 needs about 27.7M trips a year booked in the Waymo app. | ||||
| 2 | Apollo Go Driverless commercial Baidu's robotaxi app, called 萝卜快跑 in China, books nearly all of Apollo Go's driverless rides. | Driverless commercial | 4.6 | Baidu's robotaxi app, called 萝卜快跑 in China, books nearly all of Apollo Go's driverless rides. |
How we got hereHow the score was calculated10.7M fully driverless rides reported from Q3 2025 to Q2 2026, minus about 0.25M booked through AutoGo in Abu Dhabi (about 50 cars, estimated) = 10.45M → 4.62, shown as 4.6. We found no third-party app in China booking its rides; Uber in Dubai started on August 20, after this window. FootprintDriverless public rides in about 20 cities, mostly in China, plus Dubai with Dubai Taxi Company. Baidu Maps also books rides and rolls up here (same parent). Abu Dhabi rides are booked through AutoGo. What would move the scoreQ3 2026 results, due mid-November, or a third-party app in China starting to book its rides. | ||||
| 3 |  AmapDriverless commercial China's largest ride-hailing aggregator; books Pony.ai and WeRide robotaxis. | Driverless commercial | 4.4 | China's largest ride-hailing aggregator; books Pony.ai and WeRide robotaxis. |
How we got hereHow the score was calculatedNo platform publishes robotaxi rides by app. Pony.ai's and WeRide's shared fleets give half their trips to the companies' own apps; the other half is split by ride-hailing market share. Amap, with about 20% of China's ride-hailing orders (estimated), takes the largest slice: about 605 Pony.ai cars (3.0M trips) and 549 WeRide cars (2.7M), 5.8M in all (estimated) → 4.36, shown as 4.4. FootprintPony.ai and WeRide robotaxis in their Chinese service areas. What would move the scoreA published split of robotaxi orders by app. Amap's market share is the key estimate: at 10% instead of 20% it would score 4.3. | ||||
| 4 | Uber Driverless commercial The biggest third-party app for robotaxi rides: Waymo in Austin and Atlanta, WeRide and Apollo Go in the UAE. Uber also offers rides with onboard safety operators in multiple markets and has announced further partnerships. | Driverless commercial | 4.3 | The biggest third-party app for robotaxi rides: Waymo in Austin and Atlanta, WeRide and Apollo Go in the UAE. Uber also offers rides with onboard safety operators in multiple markets and has announced further partnerships. |
How we got hereHow the score was calculatedWaymo: about 400 Uber-only cars in Austin and Atlanta × 1.3 (Uber's utilization edge) × 7,362 trips per car a year = 3.8M. UAE: about 150 driverless cars × 17.5 trips a day = 1.0M. Total 4.8M (estimated) → 4.28, shown as 4.3. Rides from supervised partners (Avride, Verne, Riyadh) don't count. FootprintDriverless rides in Austin, Atlanta, Abu Dhabi and Dubai; supervised partner rides in other cities. Zoox and Motional launches are due later in November. What would move the scoreUber's UAE driverless count: at the nearly 200 it expects by year-end, its score would rise from 4.28 to 4.31, still shown as 4.3. | ||||
| 5 |  PonyPilotDriverless commercial Pony.ai's own app and WeChat mini program. | Driverless commercial | 4.3 | Pony.ai's own app and WeChat mini program. |
How we got hereHow the score was calculated1,975 robotaxis (June 30, 2026), minus about 100 dedicated to OnTime, leaves 1,875 shared across Pony.ai's booking apps. Half go to Pony.ai's own app and mini program: 937 cars × 5,000 trips a year = 4.7M (estimated) → 4.27, shown as 4.3. The other half goes to Amap, Tencent Mobility Service, Alipay Robotaxi, Sunlight Mobility and OnTime by market share. FootprintBeijing, Shanghai, Guangzhou and Shenzhen, plus newer Chinese cities. Pony.ai's rides abroad are booked through partner apps. What would move the scoreA published split of Pony.ai's orders by app. If its own apps carried three-quarters of the shared fleet instead of half, it would score 4.4. | ||||
| 6 |  WeRide GoDriverless commercial WeRide's own app and WeChat mini program in China. | Driverless commercial | 4.1 | WeRide's own app and WeChat mini program in China. |
How we got hereHow the score was calculated1,800 robotaxis (July 31, 2026), minus about 400 in the Middle East, leaves about 1,400 in China. Half go to WeRide Go: 700 cars × 5,000 trips a year = 3.5M (estimated) → 4.15, shown as 4.1. The other half goes to Amap, Tencent Mobility Service and OnTime by market share; about 1.0M UAE trips are booked in Uber. FootprintGuangzhou and Beijing, among other Chinese cities. WeRide's rides abroad are booked through Uber, Grab and other partner apps. What would move the scoreA published split of WeRide's orders by app. At three-quarters of its China fleet it would score 4.3. | ||||
| 7 |  OnTimeDriverless commercial Guangzhou ride-hailing platform that mixes human-driven cars with Pony.ai and WeRide robotaxis. | Driverless commercial | 3.9 | Guangzhou ride-hailing platform that mixes human-driven cars with Pony.ai and WeRide robotaxis. |
How we got hereHow the score was calculatedAbout 100 Pony.ai Gen-7 cars are dedicated to OnTime, which owns and dispatches them: 0.5M trips. It also gets its market-share slice (3.3% of China's ride-hailing orders) of Pony.ai's and WeRide's shared fleets: about 190 cars, 0.96M trips. Total 1.46M (estimated) → 3.86, shown as 3.9. FootprintGuangzhou's Nansha district, Shenzhen and Hengqin. What would move the scoreOnTime says it runs more than 550 robotaxis. If most are its own dedicated cars, it would score about 4.2. | ||||
| 8 | Tesla Robotaxi Driverless commercial Tesla's own app, the only way to book its driverless rides. | Driverless commercial | 3.8 | Tesla's own app, the only way to book its driverless rides. |
How we got hereHow the score was calculatedTesla reports miles, not trips. About 1.65M driverless miles in Q3 2026 × 4 ÷ 6.7 miles per trip = about 1.0M trips a year (estimated) → 3.76, shown as 3.8. The Tesla Robotaxi app is the only way to book them. Footprint6 driverless public cities; supervised service in the Bay Area; 4 cities announced. 300+ driverless vehicles. What would move the scoreReported trip counts. If October's pace holds, the score would be about 4.1. | ||||
| 9 |  Tencent Mobility ServiceDriverless commercial Ride-hailing mini program inside WeChat; books Pony.ai and WeRide robotaxis in Guangzhou. | Driverless commercial | 3.6 | Ride-hailing mini program inside WeChat; books Pony.ai and WeRide robotaxis in Guangzhou. |
How we got hereHow the score was calculatedIts market-share slice of Pony.ai's and WeRide's shared fleets, at about 2% of China's ride-hailing orders (estimated): about 66 + 60 cars = 0.63M trips (estimated) → 3.65, shown as 3.6. It also books OnTime's robotaxis, which aren't counted separately. FootprintGuangzhou, since March 13, 2026. Tencent Maps is announced next. What would move the scoreA published order figure. Its market share is estimated from the aggregators' total; at 5% it would score 3.8. | ||||
| 10 |  Sunlight MobilityDriverless commercial Chinese ride-hailing platform that funds and books Pony.ai Gen-7 robotaxis. | Driverless commercial | 3.6 | Chinese ride-hailing platform that funds and books Pony.ai Gen-7 robotaxis. |
How we got hereHow the score was calculatedSunlight funds co-built Pony.ai cars that both platforms dispatch; the number isn't disclosed. Its market-share slice of Pony.ai's shared fleet (3.3%, set equal to OnTime's) is about 101 cars = 0.50M trips (estimated) → 3.59, shown as 3.6. FootprintGuangzhou first, from late 2025; other Chinese cities planned. What would move the scoreThe number of co-built cars, or a published order figure. | ||||
| 11 | Zoox Driverless commercial Zoox's own app; the only way to book its robotaxis until the Uber launch. | Driverless commercial | 3.5 | Zoox's own app; the only way to book its robotaxis until the Uber launch. |
How we got hereHow the score was calculatedZoox reports riders, not trips. About 100 vehicles × 5,000 trips a year = 0.5M (estimated) → 3.59, which rounds to 3.6. Tencent Mobility Service and Sunlight Mobility also round to 3.6, and at most two apps share a score, so Zoox, third on full precision, shows 3.5 in this list (3.6 excluding China). All are booked in the Zoox app until its Uber launch. Footprint2 driverless public cities; about 10 testing cities; Los Angeles announced for 2027. About 100 vehicles (press reports). What would move the scoreA reported fleet or trip count. At its March–August rider pace (about 1.2M trips a year) it would score 3.8. Rides booked through Uber from late November will count for Uber. | ||||
| 12 |  Alipay RobotaxiDriverless commercial Robotaxi mini program inside Alipay, launched in February 2026; books Pony.ai and others. | Driverless commercial | 3.5 | Robotaxi mini program inside Alipay, launched in February 2026; books Pony.ai and others. |
How we got hereHow the score was calculatedIts market-share slice of Pony.ai's shared fleet, at about 2% (estimated): about 66 cars = 0.33M trips (estimated) → 3.48, shown as 3.5. FootprintParts of Guangzhou, Dongguan, Beijing, Wuhan and Shenzhen. What would move the scoreA published order figure, or confirmation of the other providers it books. Its Wuhan coverage suggests a second provider. | ||||
| 13 | AutoGo Driverless commercial Abu Dhabi app that books Apollo Go's driverless robotaxis. | Driverless commercial | 3.4 | Abu Dhabi app that books Apollo Go's driverless robotaxis. |
How we got hereHow the score was calculatedApollo Go's Abu Dhabi rides are booked through AutoGo. About 50 cars (estimated: 'dozens' at launch, 'hundreds' targeted for 2026) × 5,000 trips a year = 0.25M → 3.41, shown as 3.4. FootprintAbu Dhabi, starting on Yas Island; driverless public rides since January 2026. What would move the scoreA published fleet count. At 150 cars it would score 3.7. | ||||
| 14 | Lyft Driverless commercial Books Waymo rides in Nashville from a fleet shared with the Waymo app. | Driverless commercial | 3.3 | Books Waymo rides in Nashville from a fleet shared with the Waymo app. |
How we got hereHow the score was calculatedWaymo's Nashville fleet, 'dozens of vehicles' taken as 50, is shared by the Waymo app and Lyft, half each: 25 cars × 7,362 trips a year = 0.18M (estimated) → 3.33, shown as 3.3. May Mobility rides in Atlanta are supervised and don't count. FootprintNashville (Waymo, since September 9, 2026); supervised May Mobility rides in Atlanta. What would move the scoreA published Nashville fleet size. Lyft expects it to grow significantly by year-end; at 100 cars it would score 3.5. | ||||
| 15 | MOIA Supervised rides Volkswagen's ride-pooling app; autonomous rides in Hamburg with a safety driver. | Supervised rides | 2.2 | Volkswagen's ride-pooling app; autonomous rides in Hamburg with a safety driver. |
How we got hereHow the score was calculated1.5 for autonomous rides in the app + 0.4 for public riders (selected residents) + 0.2 for an area, not a fixed route + 0.1 for a dated driverless target (2027) = 2.2. Which app riders use isn't stated; we assume MOIA's. FootprintSelected areas of Hamburg, up to 10 vehicles, since July 2026. What would move the scoreRemoving the safety driver (+0.4), service all week for 12+ hours a day (+0.2), or 25+ vehicles (+0.1). | ||||
| 16 | Verne Supervised rides Croatian robotaxi app; paid rides in Zagreb with a safety operator, also bookable on Uber. | Supervised rides | 2.2 | Croatian robotaxi app; paid rides in Zagreb with a safety operator, also bookable on Uber. |
How we got hereHow the score was calculated1.5 for paid rides in the app + 0.4 for public riders + 0.2 for an area (city centre and airport) + 0.1 for a driverless target by end of 2026 = 2.2. Tied with MOIA, which is listed first on vehicles (10 vs not disclosed). FootprintZagreb city centre and airport, since April 2026. What would move the scoreRemoving the safety operator (+0.4), confirmed service of 12+ hours a day (+0.2), or 25+ vehicles (+0.1). | ||||
| 17 | Grab (Ai.R) Supervised rides Grab's autonomous service in Punggol, Singapore, booked through its Ai.R portal rather than the Grab app. | Supervised rides | 1.9 | Grab's autonomous service in Punggol, Singapore, booked through its Ai.R portal rather than the Grab app. |
How we got hereHow the score was calculated1.5 for autonomous rides + 0.4 for public riders = 1.9. Three fixed routes, weekdays 9:30–17:30, 10 vehicles, with a safety operator aboard. FootprintPunggol, Singapore. What would move the scoreRemoving the safety operator (+0.4), service all week for 12+ hours a day (+0.2), an area instead of fixed routes (+0.2), or 25+ vehicles (+0.1; 50 are planned by March 2027). | ||||
| 18 | Bolt Announced Testing Pony.ai robotaxis with Stellantis in Luxembourg; no rider bookings yet. Bolt also plans at least 25,000 Lucid robotaxis using NVIDIA technology in Europe; these are announced, not deployed. | Announced | 0.8 | Testing Pony.ai robotaxis with Stellantis in Luxembourg; no rider bookings yet. Bolt also plans at least 25,000 Lucid robotaxis using NVIDIA technology in Europe; these are announced, not deployed. |
How we got hereHow the score was calculated0.4 for a signed agreement naming a market (Pony.ai and Stellantis, Luxembourg) + 0.4 for vehicles deployed for the pilot = 0.8. No permit for passenger service and no dated rider launch. FootprintLuxembourg (testing). What would move the scoreA passenger-service permit (+0.3) or a dated rider launch (+0.3). Rides in the app move it to 1.5 or higher. | ||||
| 19 | inDrive Announced Agreements to bring robotaxis to Riyadh and Astana; no fleet or launch date yet. | Announced | 0.4 | Agreements to bring robotaxis to Riyadh and Astana; no fleet or launch date yet. |
How we got hereHow the score was calculated0.4 for agreements naming Riyadh (December 2025, with WeRide's Saudi dealer AI Driver) and Astana (October 2026, with Pony.ai and NVIDIA). No committed fleet, permit or dated rider launch. FootprintRiyadh and Astana (planned). What would move the scoreA committed fleet (+0.4), a permit (+0.3) or a dated launch (+0.3). | ||||
Autonomous rides only in China · Autonomous rides only or mainly in the EU
Volume unknown
Driverless rides in the app, but not enough public information to estimate trips. Not scored.
DiDi
Fully driverless ride-hailing trials, around the clock, in Guangzhou pilot zones, booked in the DiDi app since December 2025. DiDi doesn't disclose its autonomous fleet, its trips or whether riders pay.
Via
Waymo vehicles in Via's on-demand transit service in Chandler, Arizona. Volume isn't disclosed.
Methodology
How the score works
Every app gets one score from 0 to 10, shown to one decimal, rounded.
What counts
Completed autonomous rides booked in the app, whoever built the driving system or owns the car. A Waymo ride booked in Uber counts for Uber here and for Waymo in ADS Providers. A driving company's own mini programs count with its app; a third-party service that takes the booking, such as Amap or Tencent Mobility Service, counts as its own app.
Excluding and including China
The ranking is shown two ways. Excluding China, the default, counts only autonomous rides outside mainland China. Apps that operate only in China drop out, and a Chinese app that also operates elsewhere counts only those rides (Apollo Go: Dubai). Including China counts every ride. An app with no rides in China has the same trips in both; its shown score can differ only when two others share its rounded score in one list (Zoox: 3.6 excluding China, 3.5 including).
Commercial scale: 3.0 to 10
Once an app books driverless rides open to the public, its score comes from annual trips on the same scale as ADS Providers, so the two rankings compare directly.
| Score | Equivalent fleet | Trips a year |
|---|---|---|
| 3.0 | Launch (≤10 cars) | ≤50,000 |
| 4.0 | 500 | 2.5M |
| 5.0 | 5,000 | 25M |
| 6.0 | 40,000 | 200M |
| 7.0 | 200,000 | 1B |
| 8.0 | 500,000 | 2.5B |
| 9.0 | 1M | 5B |
| 10.0 | 2M | 10B, about the number of rides Uber completes worldwide each year |
Before driverless service: 0 to 2.9
- 0.0–1.4 Announced: a signed agreement or plan, no rides in the app yet.
- 1.5–2.9 Autonomous rides in the app, not yet driverless commercial service: supervised, employee-only or restricted.
- Points within a band measure how close the app is to its next stage.
Where the trips come from
Few apps report autonomous rides. We start from each driving company's annual trips, as in ADS Providers, and assign them to apps. Cars dedicated to one app go to that app. A fleet that several apps can book gives half its trips to the driving company's own app and splits the rest across the other apps by their share of ride-hailing orders. Each assessment shows the math and its inputs, with dates where the sources give them.
Volume unknown
Apps with driverless commercial rides but too little public information to estimate trips are listed below the ranking, without a score.
AVFox distinguishes announcements, permits, testing, and commercial operation. Coverage does not imply that a company has deployed or operates commercially in every market.