Robotaxis: which startup is ahead?
In our autonomous vehicle market deck, you will find everything you need to understand the market
SUMMARY
Pony.ai is the robotaxi startup ahead today, with the largest independent fleet, the fastest recent growth in paid robotaxi revenue and the clearest evidence that mass production is improving vehicle economics.
WeRide is much closer than the rest of the field. Its fleet is about three-quarters the size of Pony.ai’s, it generated more robotaxi revenue during 2025, and it has taken fully driverless paid service into more international markets.
The race has split into two tiers. Pony.ai and WeRide operate roughly 3,000 robotaxis between them, while the next disclosed fleet belongs to Avride and includes only a little more than 200 commercial and research vehicles.
Commercial scale and technical maturity are no longer the same thing. Avride has completed more than 60,000 paid Uber trips but still uses onboard specialists, while May Mobility runs smaller services and has already removed the driver in selected locations.
Pony.ai’s strongest recent advantage is not fleet size alone. Its latest quarter produced $8.6 million in robotaxi revenue, paid orders increased sharply, and the company says its newest vehicles have reached positive unit economics in Guangzhou and Shenzhen.
WeRide has built the more international operating model. Uber and Grab bring passengers and payments, local partners help manage fleets, and WeRide focuses on the driving system, vehicle integration and regulatory approvals.
Nuro and Wayve are serious challengers, but their evidence is still mostly pre-launch. Nuro has nearly 100 Lucid engineering vehicles and a 20,000-vehicle commitment, while Wayve has major automaker partnerships but no large public driverless fleet.
Manufacturing could decide the next phase. Pony.ai is already deploying its seventh-generation system through Toyota, BAIC and GAC, while WeRide is preparing a 2,000-vehicle program with Geely Farizon.
Safety remains the least defensible part of any ranking. The startups do not publish comparable driverless mileage, crash and intervention data, so permits and operating history can show regulatory maturity but cannot prove which company is safest.
The ranking can still change quickly. WeRide could take first place if its new fleet arrives on schedule and fills through Uber and Grab, while a clean US launch could move Nuro above May Mobility and Avride almost immediately.
For now, the order is Pony.ai, WeRide, May Mobility, Avride, Nuro, Wayve and Momenta. The gap between first and second is modest; the gap between second and third is large.
This market map, featured in our autonomous vehicle market deck, highlights top companies and startups in the autonomous vehicle market
Which robotaxi startups are serious enough to compare?
The robotaxi startup race currently has seven credible contenders, although Pony.ai and WeRide have already pulled far ahead in commercial scale.
We include venture-backed autonomous-driving companies that remain independent from a large technology or automotive parent. Pony.ai, WeRide and Momenta still qualify even though they have gone public because they were built as standalone autonomous-driving challengers.
Waymo, Zoox, Cruise, Motional and Apollo Go sit outside our scope. Alphabet controls Waymo, Amazon owns Zoox, General Motors controls Cruise, Hyundai owns Motional, and Apollo Go belongs to Baidu. Tesla is also excluded because its robotaxi program sits inside a much larger automaker.
Those exclusions change the answer. Waymo is clearly ahead across the whole robotaxi industry, with Alphabet reporting more than 500,000 fully autonomous rides a week. The question here is which startup has come closest to building a comparable business.
We also leave out companies whose recent scale is difficult to verify. An old testing permit or a polished demonstration no longer earns a place among the serious contenders. We want evidence of vehicles operating today, paying passengers, recent production, or a credible path to public service.
Funding figures remain approximate. Private rounds are disclosed unevenly, IPO proceeds are different from venture funding, and Avride’s headline amount includes commercial commitments as well as investment.
| Startup | What the company is building | Where its robotaxi business stands today | Approximate disclosed capital |
|---|---|---|---|
| Pony.ai | Full-stack urban robotaxis using vehicles from Toyota, BAIC and GAC | More than 1,700 robotaxis, paid services and fully driverless operations | About $2.5 billion |
| WeRide | Robotaxis alongside autonomous buses, vans and street sweepers | Around 1,300 robotaxis, with fully driverless paid services in China and the UAE | About $1.6 billion |
| May Mobility | Autonomous ride-hailing and public transport in defined operating areas | More than 500,000 rides across 24 deployments, with some driverless operation | About $390 million |
| Avride | Robotaxis and delivery robots distributed through Uber | More than 60,000 paid Dallas trips and over 200 vehicles, including R&D vehicles | Up to $375 million in investment and commitments |
| Nuro | Autonomous-driving software for passenger vehicles | Nearly 100 Lucid engineering vehicles ahead of a planned Uber launch | More than $3 billion |
| Wayve | End-to-end driving AI for automakers and robotaxi networks | Large international testing footprint, with commercial trials beginning soon | About $2.6 billion |
| Momenta | Assisted-driving software and a developing Level 4 robotaxi service | Large automaker footprint, while robotaxi deployments remain early | About $2 billion, including public financing |
Is Pony.ai already the clear robotaxi startup leader?
Pony.ai is currently the strongest robotaxi startup overall, but WeRide remains close enough to take the lead if its international services scale faster.
The competitive structure is fairly clear now. Pony.ai and WeRide form the top group. May Mobility and Avride have genuine passenger activity at a much smaller scale. Nuro and Wayve are well-funded challengers preparing for commercial launches. Momenta’s strongest business still lies in production-car software rather than robotaxis.
Pony.ai and WeRide together operate roughly 3,000 robotaxis. Avride, the next company with a disclosed fleet above 200 vehicles, remains several times smaller. Nuro has nearly 100 engineering vehicles, while Wayve has yet to publish a comparable operating fleet.
Pony.ai gets the overall lead from the combination of fleet size, paid-order growth and progress toward positive vehicle economics. WeRide counters with higher robotaxi revenue during 2025, a broader international presence and fully driverless Uber rides in the Middle East.
Neither company has built a mature financial business yet. Pony.ai reported a quarterly operating loss of roughly $58 million in its latest results, while WeRide lost about $63 million from operations. Even so, the operating gap looks closer to 100 versus 80 than 100 versus 20.
If you want more recent data on this point, please see our latest autonomous vehicle market report.
As this chart shows, and as featured in our autonomous vehicle market deck, search interest in autonomous vehicles has continued to rise
Which robotaxi startup has the largest fleet today?
Pony.ai currently has the largest disclosed startup robotaxi fleet, with more than 1,700 vehicles compared with around 1,300 for WeRide.
Pony.ai’s fleet is roughly 31% larger. Put another way, WeRide operates about 76 robotaxis for every 100 operated by Pony.ai. The comparison becomes much less competitive after those two companies.
Avride says its total fleet exceeds 200 vehicles, although that figure combines commercial robotaxis with research and development vehicles. Nuro has nearly 100 Lucid Gravity engineering vehicles. May Mobility has operated across many locations but does not publish one current global fleet total. Wayve and Momenta also keep their fleet numbers private.
Recent additions show that Pony.ai and WeRide are still accelerating. Pony.ai ended 2025 with more than 1,400 robotaxis and has since moved beyond 1,700. WeRide grew from 1,023 vehicles near the beginning of the year to approximately 1,300 by the end of April.
Pony.ai has raised its year-end target above 3,500 vehicles, while WeRide expects more than 2,600 to be operating. Those targets remain production plans until the vehicles are built, licensed and placed into active service.
| Startup | Latest disclosed fleet | What the figure includes | Fleet index, Pony.ai = 100 |
|---|---|---|---|
| Pony.ai | More than 1,700 | Deployed robotaxis | 100 |
| WeRide | About 1,300 | Global robotaxi fleet | 76 |
| Avride | More than 200 | Commercial and R&D vehicles | At least 12 |
| Nuro | Nearly 100 | Lucid engineering vehicles | About 6 |
| May Mobility | Not centrally disclosed | Vehicles across several local deployments | Not comparable |
| Wayve | Not disclosed | Test vehicles and upcoming pilot fleets | Not comparable |
| Momenta | Not disclosed | Test and development vehicles | Not comparable |
Which robotaxi startup has built the biggest paid business?
WeRide built the larger robotaxi business during 2025, while Pony.ai now appears to have the stronger revenue trajectory.
WeRide reported $21.2 million in robotaxi revenue for 2025, up 210% from the previous year. Pony.ai generated $16.6 million, representing 129% growth. On that full-year comparison, WeRide earned about 28% more.
Pony.ai then accelerated sharply. Its robotaxi business generated $8.6 million in the latest quarter, more than half of its entire 2025 total in only three months. Fare-charging revenue rose more than fivefold from the same quarter a year earlier.
WeRide did not publish a clean robotaxi-only number for its latest quarter. Total company revenue reached $16.5 million, including vehicle sales, data services, assisted-driving work and other activities. We cannot claim that Pony.ai has already overtaken WeRide without a comparable quarterly figure.
Avride has completed more than 60,000 paid Uber trips in Dallas but has not disclosed the revenue earned from them. May Mobility reports more than half a million cumulative rides, although that total combines several commercial and public-sector models. Nuro and Wayve still have no broad public robotaxi revenue.
WeRide therefore holds the stronger completed annual result. Pony.ai is closing the gap fastest.
This chart, included in our autonomous vehicle market deck, illustrates yearly VC funding for autonomous vehicle startups
Which robotaxi startup is growing fastest right now?
Pony.ai is currently growing fastest where it counts most: paying users, completed orders and robotaxi revenue.
Pony.ai’s latest quarterly report showed robotaxi revenue increasing 395% from a year earlier. Its registered user base in China more than tripled, while average weekly paid orders increased 119% between January and May. The three measures are moving together rather than reflecting fleet growth alone.
WeRide is also growing quickly. Its robotaxi fleet expanded by roughly 27% between January and the end of April. Average daily orders exceeded 17 per vehicle during the first quarter and reached 28 at peak times. Registered users roughly doubled from a year earlier.
Avride has produced one of the freshest signs of demand in the smaller group. The company passed 60,000 Uber passenger trips in Dallas, doubled its original service area and reported 1.3 million autonomous miles across its fleet. One million of those miles were driven during 2026.
May Mobility is expanding through partners rather than through one rapidly growing consumer network. The company recently announced a European exploration agreement with CaoCao, a fleet-scaling partnership with ECARX and upcoming Uber operations in Arlington. Most of those projects have yet to produce large ride volumes.
Wayve has attracted new capital and signed Uber partnerships involving Nissan and Stellantis. Nuro has secured permits, prepared nearly 100 vehicles and added Houston as its second planned market after the San Francisco Bay Area. Their progress is still measured through preparation rather than public rides.
Pony.ai leads the current growth comparison. WeRide is closest, and Avride has the most interesting momentum outside the top two.
If you want more recent data on this point, please see our latest autonomous vehicle market report.
Which robotaxi startups are truly driverless today?
Pony.ai and WeRide are currently the only startups operating fully driverless, fare-charging robotaxis at meaningful urban scale.
Pony.ai has removed the onboard safety driver from commercial services in several Chinese cities. WeRide has done the same in China, Abu Dhabi and Dubai. Riders in parts of the UAE can request a WeRide vehicle through Uber and complete the trip without anyone sitting behind the wheel.
May Mobility has run driverless vehicles in selected US locations, including smaller and more controlled service areas. The company recently referred to its second US city with driver-out operations. Its scale remains far below Pony.ai and WeRide.
Avride’s Dallas service still uses an onboard specialist. The 60,000 paid trips prove demand, but the service has yet to show that it can operate without onboard labor.
Nuro has secured California permits for driverless testing and for a passenger pilot, while selected Uber employees have started using its service. Wayve’s Tokyo pilot and Momenta’s planned Munich tests remain in the validation stage.
Pony.ai leads on deployed driverless vehicles. WeRide has done the better job of taking fully driverless service into several countries.
This chart, included in our autonomous vehicle market deck, shows how Waymo is winning in autonomous vehicles
Which robotaxi startup is winning outside its home market?
WeRide is currently the most international robotaxi startup, with a wider mix of permits, paying passengers and local partners than Pony.ai.
WeRide says its autonomous vehicles have been tested or operated in more than 40 cities across 12 countries. It has obtained autonomous-driving permits in eight markets: China, the UAE, Singapore, France, Switzerland, Saudi Arabia, Belgium and the United States.
The quality of those overseas operations matters more than the country count. WeRide already offers fully driverless Uber rides in Abu Dhabi and Dubai. Its Singapore service uses Grab, giving the company another established ride-hailing channel.
WeRide and Uber also plan to deploy 1,200 robotaxis across Abu Dhabi, Dubai and Riyadh by 2027. The target is ambitious, but the companies are expanding from existing paid services and regulatory approvals.
Pony.ai has established a presence in nine countries and started public services in Croatia, Qatar, Singapore and South Korea. Its Croatian deployment with Verne has begun generating project revenue, and the company is preparing driverless operations in Dubai.
May Mobility has operated in Japan, received investment from Grab for Southeast Asian expansion and recently began exploring European deployments with CaoCao. These plans remain earlier than WeRide’s live overseas services.
WeRide leads internationally by a comfortable margin. Pony.ai is the only startup close enough to challenge it.
Which robotaxi startup can manufacture at scale and make the economics work?
Pony.ai currently has the strongest combination of large-scale production, falling vehicle costs and improving per-car economics.
Pony.ai’s seventh-generation platform is already being deployed across vehicles from Toyota, BAIC and GAC. The company has secured 1,000 Toyota bZ4X-based robotaxis and is pushing the total cost of its domestic vehicle and autonomous-driving hardware below RMB230,000 by the middle of 2027.
That target would equal roughly $32,000 at recent exchange rates. The exact level has yet to be reached, but Pony.ai is already deploying the vehicle generation on which the cost claim is based.
The company also says its seventh-generation robotaxis have reached positive unit economics in Guangzhou and Shenzhen. In Shenzhen, each vehicle averaged 23 paid orders and RMB338 in daily net revenue during February. Peak performance later reached 25 orders and RMB394.
Pony.ai does not publish a full cost breakdown covering depreciation, remote assistance, cleaning, charging, insurance and maintenance. The claim shows progress at the vehicle level rather than proof that the whole company is profitable.
WeRide is taking a more concentrated production route with Geely Farizon. The companies plan to deliver 2,000 upgraded GXR robotaxis, with production expected to begin during the third quarter. WeRide says the autonomous hardware costs about $40,000 per vehicle and could fall by another 15% as production grows.
WeRide reported a 34.7% company-wide gross margin in its latest quarter, compared with Pony.ai’s 16.2%. The companies sell different mixes of vehicles and services, so the figures do not reveal which robotaxi earns more per ride.
Funding remains another part of the comparison. Pony.ai has raised roughly $2.5 billion and reported more than $1.5 billion in cash reserves after its Hong Kong listing. WeRide has raised closer to $1.6 billion and ended its latest quarter with about $902 million in cash, time deposits and similar assets.
Nuro holds the largest long-term production commitment through Uber and Lucid, covering at least 20,000 vehicles over six years. Nearly 100 engineering vehicles are active today, so that program remains a future scale opportunity rather than current manufacturing leadership.
Pony.ai stays ahead because it has already connected production volume with the largest startup fleet and improving vehicle economics. WeRide appears more capital-efficient and could close the gap quickly if its 2,000-vehicle program stays on schedule.
If you want more recent data on this point, please see our latest autonomous vehicle market report.
This chart, included in our autonomous vehicle market deck, illustrates yearly funding for autonomous vehicle startups
What can Pony.ai and WeRide do that rivals cannot copy quickly?
Pony.ai and WeRide have built an operating loop that a well-funded newcomer cannot reproduce simply by buying more sensors and hiring more engineers.
Thousands of active vehicles expose the software to unusual road situations every day. Engineers use that experience to improve the system. The improved system supports stronger safety cases, which help win new permits. New permits produce more rides, revenue and operating data.
Pony.ai says its shared robotaxi and robotruck technology has accumulated more than 70 million autonomous kilometres. Because that total includes trucks, it should not be treated as pure robotaxi mileage. It still reflects years of experience across several vehicle types and demanding road conditions.
WeRide’s regulatory history creates a different advantage. The company has persuaded authorities in eight countries to grant some form of autonomous-driving permit. Each market has different road rules, vehicle standards, data requirements and approval processes.
Its overseas services have also taught WeRide how to divide responsibilities with partners. Uber or Grab can supply passengers and payments. Local transport companies handle parts of fleet management. WeRide concentrates on the driving system, calibration and regulatory work.
Pony.ai has developed similar knowledge through OnTime Mobility, Verne and several automakers. Its ability to integrate the same autonomous-driving platform into vehicles from three manufacturers reduces the risk of being trapped by one supplier.
Wayve hopes to weaken this advantage through an end-to-end AI system that adapts more easily to unfamiliar cities. The company says its vehicles drove “zero-shot” in more than 500 cities within a year, meaning the software entered those places without city-specific retraining.
That is impressive, although supervised testing across many cities differs from running thousands of unattended rides in the same cities every week. Wayve now has to prove that broad adaptability can reach the reliability demanded by a public robotaxi service.
Is Wayve already a real robotaxi contender?
Wayve is a serious technology contender but remains outside the leading robotaxi group because paying passengers have barely entered the picture.
The company has raised about $2.6 billion, including a recent $1.5 billion financing package. Uber will provide part of that capital as Wayve reaches deployment milestones.
Wayve’s appeal comes from its end-to-end AI approach. Traditional robotaxi systems often combine hand-designed rules, high-definition maps and separate software modules for perception, prediction and planning. Wayve trains a more unified model to learn driving behaviour across vehicles and locations.
The company has assembled several credible routes to market. Nissan and Uber plan a Tokyo pilot using a Nissan Leaf equipped with Wayve software. Stellantis, Wayve and Uber have also agreed to explore Level 4 robotaxis globally, adding another large vehicle supplier.
Wayve is preparing trials in London, Tokyo and other cities, while its supervised-driving technology should reach consumer vehicles from 2027. These partnerships make it more than a laboratory project.
The missing evidence is still substantial. Wayve does not disclose a large operating robotaxi fleet, paid-ride volume, driverless commercial revenue or prolonged unattended operation. The upcoming pilots should answer some of those questions.
Wayve could become one of the biggest winners if its software works across many automakers without years of rebuilding for every city. Today, companies already carrying passengers every day rank higher.
If you want more recent data on this point, please see our latest autonomous vehicle market report.
This chart, included in our autonomous vehicle market deck, compares the main business model options for autonomous trucking companies
Can Nuro jump ahead with Uber and Lucid?
Nuro has the strongest chance of jumping from pre-launch status into the leading group, especially in the United States.
The Nuro, Uber and Lucid partnership gives each company a clear job. Lucid supplies the Gravity electric vehicle. Nuro provides the Level 4 driving system. Uber brings passenger demand, fleet tools and the customer app.
The partners intend to deploy at least 20,000 vehicles over six years across American and international markets. The San Francisco Bay Area should receive the first service, with Houston planned as the second market during 2027.
Nuro is further along than a normal future announcement suggests. Nearly 100 Lucid engineering vehicles are already supporting testing in several states. Selected Uber employees began requesting the vehicles through the Uber app in the Bay Area, and Nuro has obtained California permits covering driverless testing and a passenger pilot.
The Houston plans add physical preparation. Uber has secured a 50,000-square-foot depot and another support site for cleaning, charging and maintaining the future fleet.
Nuro still faces a large distance to the leaders. One hundred engineering vehicles represent roughly 6% of Pony.ai’s deployed fleet. Employee testing cannot tell us how regular passengers will react, how many rides each vehicle will complete or how often remote assistance will be needed.
The planned launch will be decisive. Several hundred public, fully driverless vehicles with good utilization would move Nuro above May Mobility and Avride quickly. Delays or prolonged safety-operator use would leave Nuro in the challenger group despite its funding.
Are May Mobility and Avride closer to the robotaxi leaders than they look?
May Mobility and Avride have more commercial substance than many better-funded startups, although their current robotaxi services remain much smaller than Pony.ai’s and WeRide’s.
May Mobility has completed more than 500,000 rides across 24 deployments. The company has operated in large cities, suburban communities, university areas, smaller towns and locations with difficult winter weather.
May Mobility chooses defined service areas instead of trying to open an entire city immediately. That approach creates smaller fleets but solves specific transport problems sooner. Cities can use the vehicles to connect neighbourhoods with public transit, serve people without cars or extend service into routes that conventional buses handle poorly.
The company has also started moving toward broader ride-hailing. Lyft distributes its Atlanta service, while Uber plans an Arlington deployment before expanding into additional US markets. Grab, CaoCao and ECARX give May Mobility possible routes into Asia, Europe and larger fleets.
Avride follows a more conventional robotaxi path. Dallas passengers can receive one of its Hyundai Ioniq 5 vehicles through the normal Uber app. The service has passed 60,000 trips, its operating area has doubled, and the wider fleet has travelled more than 1.3 million autonomous miles.
Avride’s current limitation sits in the front seat. An onboard specialist still monitors the Dallas rides. May Mobility has already demonstrated driver-out operations in selected locations, giving it the higher level of autonomous maturity.
Avride may be scaling passenger demand faster. The Uber integration brings regular riders without asking them to download a separate service, and the company recorded most of its autonomous mileage within a relatively short recent period.
We place May Mobility third because of its longer record, cumulative rides and driverless operations. Avride ranks fourth, with a realistic chance of overtaking May Mobility once it removes the onboard specialists and expands beyond one commercial city.
This chart, featured in our autonomous vehicle market deck, shows the share of revenue generated by each customer segment in the autonomous vehicle market
Can we honestly rank robotaxi safety?
We cannot honestly name the safest robotaxi startup today because Pony.ai, WeRide and their rivals do not publish comparable independent safety data.
The ideal comparison would show driverless miles, crashes per million miles, injury severity, responsibility for each collision and matched human-driving results from the same streets and hours.
Public disclosures fall well short of that standard. Pony.ai and WeRide publish large mileage totals, regulatory approvals and descriptions of their safety systems. The companies provide less detail on standardized crash frequency and intervention rates.
A permit still tells us something useful. Regulators in several major cities have reviewed the technology closely enough to allow paying passengers to travel without a driver. WeRide’s approvals across several countries are particularly persuasive evidence of regulatory maturity.
May Mobility operates on narrower routes, which makes a direct comparison unfair. Avride’s onboard specialists change its risk profile. Nuro, Wayve and Momenta have too little public driverless passenger exposure for a meaningful statistical comparison.
Waymo offers a useful benchmark for disclosure. A peer-reviewed study covering 7.14 million rider-only Waymo miles found an 80% lower injury-reported crash rate and a 55% lower police-reported crash rate than matched human benchmarks. The researchers still warned about reporting differences and statistical limitations.
None of the startups currently offers a similarly robust public dataset. We can rank their progress toward fully driverless service, but a safety league table would pretend the evidence is cleaner than it is.
Which robotaxi startups are actually ahead?
Pony.ai is the robotaxi startup ahead overall today, followed closely by WeRide and then by a much smaller second group led by May Mobility.
Pony.ai wins because the strongest pieces fit together. It operates the largest startup fleet, has rapidly increasing paid orders, generated $8.6 million in quarterly robotaxi revenue and reports positive unit economics in two major Chinese cities. Its seventh-generation vehicles are already being produced through several automakers.
WeRide remains a close second. It earned more robotaxi revenue during 2025, operates roughly three-quarters as many vehicles as Pony.ai and leads international deployment. No other startup currently offers fully driverless paid services across such a varied group of countries.
The gap between WeRide and third place is much wider than the gap between first and second. May Mobility has completed plenty of real rides and removed drivers in selected locations, but its fleets and operating areas are smaller. Avride is growing quickly through Uber, although its Dallas service still uses onboard specialists.
Nuro ranks fifth because the Uber and Lucid program could produce a large US fleet much faster than an independent launch. The company now has permits, vehicles, depots and two identified markets. Public service will show how much of that preparation converts into real scale.
Wayve ranks sixth. Its technology may eventually travel between vehicles and cities more easily than today’s systems, and its partnerships have become much stronger lately. The absence of a large public driverless service keeps it below Nuro.
Momenta sits seventh within robotaxis. The company has deep automaker relationships and production experience, but most of that strength currently serves assisted-driving systems. Its planned Uber tests in Munich could begin changing the picture.
Pony.ai’s lead can disappear. WeRide could move into first place by deploying its new GXR fleet on schedule and filling those vehicles through Uber and Grab. Nuro could enter the top three after a successful driverless US launch. Wayve could rise even faster if its first pilots prove that one AI system can work across several automakers and cities.
The answer today is Pony.ai. The company is ahead by a noticeable margin, while WeRide remains close enough that the next production cycle could reverse the order.
| Rank | Startup | Why it ranks here |
|---|---|---|
| 1 | Pony.ai | Largest startup fleet, strongest recent paid growth, improving vehicle economics and three major manufacturing partners |
| 2 | WeRide | Closest overall challenger, with stronger international reach and more robotaxi revenue during 2025 |
| 3 | May Mobility | More than 500,000 rides, repeated deployments and driverless operation in selected US locations |
| 4 | Avride | Fast-growing Uber service with more than 60,000 Dallas trips, but onboard specialists remain |
| 5 | Nuro | Best-funded near-term US challenger, backed by Uber, Lucid, nearly 100 engineering vehicles and a 20,000-vehicle plan |
| 6 | Wayve | Strong funding, adaptable AI and major automaker partnerships, without a large public driverless fleet yet |
| 7 | Momenta | Extensive automaker adoption and strong driving software, while its robotaxi business remains secondary |
If you want more recent data on this point, please see our latest autonomous vehicle market report.
This chart, included in our autonomous vehicle market deck, shows how robotaxi platform technology has evolved over time
OUR METHODOLOGY
This analysis asks which independent robotaxi startup is ahead based on operating scale, paying passenger activity, driverless deployment, robotaxi revenue, international expansion, manufacturing readiness and vehicle economics. The final ranking combines those dimensions rather than allowing one fleet number, funding round or partnership announcement to decide the result.
We limited the field to venture-backed autonomous-driving companies built as standalone challengers. Pony.ai, WeRide and Momenta remain in scope despite their public listings. Waymo, Zoox, Cruise, Motional, Apollo Go and Tesla are excluded because their robotaxi programs sit inside much larger technology or automotive groups.
For each question, we used the freshest operating and financial evidence available. Vehicles already deployed count more than production targets, fare-charging rides count more than demonstrations, fully driverless services count more than supervised tests, and reported revenue counts more than estimated demand.
The final order is a structured judgement rather than a mechanical score. A company ranks higher when several facts support the same conclusion: a large fleet is producing paid orders, those orders are producing revenue, manufacturing can add more vehicles, and permits or partnerships provide a credible route into additional cities.
Company disclosures are not standardized. Pony.ai and WeRide publish fleet and financial figures, May Mobility emphasizes cumulative rides and deployments, Avride reports trips and autonomous mileage, and Nuro and Wayve mainly disclose engineering fleets, permits, partnerships and launch plans. We compare those figures only within the questions they can genuinely answer.
Safety is treated separately because the startups do not publish comparable independent datasets covering driverless miles, crash severity, responsibility and intervention rates. We use permits and driverless operating history as evidence of regulatory maturity, but not as proof that one company is statistically safer than another.
Key sources include Pony.ai’s first-quarter 2026 results, Pony.ai’s full-year 2025 results, Pony.ai’s Shenzhen unit-economics update, WeRide’s latest fleet and financial update, and WeRide’s full-year 2025 results.
Additional operating sources include May Mobility’s investor information, Uber’s Avride launch announcement, Nuro’s Houston launch update, the original Uber, Nuro and Lucid partnership, Wayve’s latest financing and deployment update, and the Wayve, Nissan and Uber Tokyo pilot announcement.
For the safety-disclosure benchmark, we used Waymo’s peer-reviewed comparison of rider-only crash data with human benchmarks and Waymo’s current safety-impact methodology. These sources do not place Waymo inside the startup ranking; they show what a more comparable public safety dataset can look like.
In our autonomous vehicle market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
