Self-driving software: which startup is ahead?
In our autonomous vehicle market deck, you will find everything you need to understand the market
SUMMARY
Momenta is ahead overall in self-driving software today, with Aurora leading the harder Level 4 public-road race and Kodiak providing the clearest proof that a customer can own driverless vehicles and put them to work.
The field has split into three different businesses. Momenta sells supervised-driving software at automotive scale, Aurora sells driverless trucking capability on public highways, and Kodiak sells a narrower system that is already doing paid industrial work.
Momenta’s advantage is commercial rather than purely technical. More than 733,000 production installations, revenue of about $335 million in 2025 and rapidly expanding licensing sales put it in a different business category from rivals still funded mainly by investors.
Aurora’s advantage is harder to reproduce in the near term. Nearly 440,000 driverless public-road miles across ten routes show that its trucks can operate without anyone in the cab, although the company still needs much more revenue and many more active vehicles.
Kodiak has found a practical shortcut to commercialization. Atlas owns and operates 28 driverless trucks in the Permian Basin, so the software is already embedded in a customer’s daily operation rather than sitting inside a vendor-run pilot.
The private challengers have stronger pipelines than deployments. Wayve has the broadest automaker interest, Nuro has the most concrete new robotaxi program, Waabi has exceptional funding and simulation claims, and PlusAI has real supervised freight but no driverless launch yet.
Capital raised is a poor guide to present leadership. Wayve, Nuro and Waabi have collectively attracted billions of dollars, yet Kodiak has reached paid customer-owned driverless work with far less cumulative funding.
The most defensible moat is becoming production and operating data, not a particular AI architecture. Momenta has repeated vehicle integration across 68 production models, while Aurora and Kodiak are accumulating the maintenance, remote-support and safety experience needed to run driverless fleets.
Performance claims remain difficult to compare. Public-road miles, paid industrial hours, production kilometres, city counts and simulated coverage each prove something useful, but none creates a universal safety ranking.
Momenta keeps first place unless the comparison is restricted to fully driverless Level 4 operation. Aurora can overtake it by turning technical leadership into hundreds of revenue-generating trucks, while Kodiak can close the gap by moving its customer-owned model from industrial routes onto open highways.
This market map, featured in our autonomous vehicle market deck, highlights top companies and startups in the autonomous vehicle market
Which self-driving software startups should we compare?
We are comparing eight companies that sell or license the software driver itself: Momenta, Aurora, Kodiak, Wayve, Nuro, PlusAI, Waabi and Oxa.
The boundary is important. Waymo and Zoox operate major autonomous fleets, but they belong to Alphabet and Amazon. Tesla, Mobileye and Nvidia are established public companies with much broader businesses. Cruise is controlled by General Motors. Including them would turn this into an article about the entire autonomous-vehicle industry rather than a comparison of startup-built software platforms.
We also leave out companies whose main product is a transport service. May Mobility runs autonomous transit. Gatik moves freight on fixed middle-mile routes. Einride combines vehicles, freight operations and software. Their technology belongs in the wider market, but it is difficult to isolate the value and maturity of the underlying driver.
Pony.ai and WeRide present a similar problem. Both develop substantial autonomy software, yet their businesses are built mainly around operating robotaxis and autonomous services. This article focuses more narrowly on companies trying to place one software platform inside vehicles made, owned or operated by other businesses.
Aurora, Momenta and Kodiak are now publicly traded, but all three grew out of the same venture-backed startup race as Wayve, Nuro and Waabi. We therefore keep them in the field while counting identifiable public financing alongside private funding. The capital figures below are approximate because companies disclose rounds differently, and milestone-based commitments are not the same as cash already received.
| Startup | What it sells | Approximate cumulative capital | Best proof available today |
|---|---|---|---|
| Momenta | Automated-driving and future robotaxi software for passenger cars | About $1.8 billion including its IPO | More than 733,000 production installations |
| Aurora | Driver-as-a-service software for autonomous trucks | More than $3 billion in private and public capital | Nearly 440,000 driverless public-road miles |
| Kodiak | Autonomous software for industrial, freight and defense vehicles | About $540 million including its latest PIPE | 28 customer-owned driverless vehicles and 23,500 paid hours |
| Wayve | End-to-end AI driving software for passenger cars and robotaxis | About $2.5 billion closed, potentially $2.8 billion with milestone capital | Testing in more than 500 cities and several automaker programs |
| Nuro | Licensed autonomy for robotaxis and passenger vehicles | More than $2.3 billion | A production-intent Lucid robotaxi and nearly 100 engineering vehicles |
| Waabi | Simulation-led autonomy for trucks and robotaxis | About $1.03 billion closed, potentially $1.28 billion with milestones | Volvo truck integration and a large proposed Uber program |
| PlusAI | Factory-integrated autonomous-trucking software | About $720 million | A factory-integrated truck moving freight on a daily 600-mile route |
| Oxa | Configurable autonomy for industrial vehicles | About $349 million | Live deployments in airports, ports and industrial sites |
Is there a clear self-driving software leader today?
Today, Momenta leads the commercial software business, Aurora leads driverless operation on public roads, and Kodiak leads customer-owned driverless work.
Momenta has reached a scale normally associated with a mature automotive supplier, but most cars using its software still require an attentive driver. Aurora operates far fewer vehicles, yet its trucks can travel on public highways with nobody in the cab. Kodiak works in a narrower environment, although its customer already owns the trucks and uses them for paid work.
Wayve sits just behind those three. Several major automakers now want to place its AI Driver inside future vehicles, and Uber is involved in more than one planned robotaxi deployment. Wayve currently has an impressive collection of partners and tests, but very few ordinary customers can use its software.
Nuro is moving closer to launch through its work with Lucid and Uber. PlusAI is carrying real freight under supervision. Waabi has attracted extraordinary capital and serious industrial partners. Oxa has working deployments in industrial environments.
Momenta is currently the strongest business. Aurora has built the strongest Level 4 product. Kodiak has gone furthest in proving that a customer can own autonomous vehicles and put them to work.
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 startup has built the biggest self-driving software business?
Momenta has built the biggest self-driving software business by a wide margin.
Momenta’s Hong Kong listing prospectus gives us unusually detailed financial and operating data. Revenue reached RMB2.41 billion in 2025, around $335 million, after growing 78.4% in 2024 and another 82.1% in 2025. None of the other companies in this comparison reports autonomy revenue in the same order of magnitude.
The composition of that revenue has also changed. Licensing sales increased from RMB23 million in 2023 to RMB968 million two years later. Licensing consequently grew from 3.1% to 40.1% of total revenue. Momenta is gradually moving away from depending mainly on engineering projects and toward earning money each time an automaker produces another equipped vehicle.
Gross margin followed the same trajectory, rising from 17.5% in 2023 to 71.6% in 2025. Revenue added more than RMB1 billion during the latest year, while the direct cost of generating that revenue barely moved. That is the closest any company in this field has come to showing genuine software economics.
Vehicle deployment expanded just as quickly. According to Momenta’s prospectus, more than 733,000 production vehicles had been installed with its systems by the latest reporting cut-off. The company had secured nominations for 180 models, of which 68 had entered production.
Aurora recorded only about $3 million of revenue in 2025. Kodiak’s quarterly revenue recently passed $1 million. The private companies do not publish comparable figures, and several remain almost entirely dependent on investment capital.
Momenta still loses money and depends heavily on China. Its installed base also consists mainly of supervised driving systems rather than robotaxis. That is a big caveat. Still, no rival currently comes close to matching its combination of vehicle volume, annual revenue, licensing growth and gross margin.
Which self-driving startup is running vehicles with nobody behind the wheel?
Aurora currently has the strongest public-road driverless operation, while Kodiak has the strongest paid driverless operation on private industrial routes.
Aurora’s latest company disclosure puts its trucks at nearly 440,000 miles without anyone behind the wheel. They now operate across ten routes in the US Sun Belt, including high-speed freight corridors rather than a small closed campus. Seven customers had joined Aurora’s driverless group by its latest quarterly update.
Aurora has also started deploying its second-generation truck. The new International LT platform runs without the observer previously requested by a partner, uses hardware designed to last one million miles and has passed Aurora’s internal safety-case process. Roush installs the redundant systems and autonomy hardware at a facility dedicated to Aurora.
Kodiak has progressed further with customer ownership. Its first-quarter filing reported 28 customer-owned driverless vehicles, up from 20 at the end of the previous quarter. Those trucks had completed more than 23,500 paid driverless hours and delivered over 15,600 loads.
Atlas Energy Solutions uses the vehicles to transport sand in the Permian Basin. Nobody sits inside the cab, and Atlas rather than Kodiak owns and operates the fleet. The routes are less complicated than Aurora’s public highways, but the commercial arrangement is already part of Atlas’s daily business.
Nuro has genuine driverless experience from its earlier delivery business, although its new passenger platform remains in testing. Wayve, PlusAI and Waabi still rely on supervised roads, closed courses or forthcoming pilots for their strongest demonstrations.
This chart, included in our autonomous vehicle market deck, illustrates yearly VC funding for autonomous vehicle startups
Which self-driving software startup is growing fastest now?
Momenta is growing fastest as a software business, while Kodiak is expanding fastest from a small driverless operating base.
Momenta’s revenue has grown by more than 78% in each of the past two reported years. Licensing sales increased more than fortyfold between 2023 and 2025, and its number of production models rose from eight to 68. Several measures moved together, which makes the growth much harder to dismiss as the result of one unusually large project.
Kodiak’s latest quarter produced the most dramatic driverless growth. Its customer-owned fleet increased 40%, from 20 to 28 vehicles. Paid hours rose 120%, considerably faster than the fleet itself, suggesting that the existing trucks were also being used more intensively. Cumulative loads increased by about 24%.
Aurora’s recent growth is better measured through operating coverage than revenue. Its driverless network reached ten routes, McLane joined its customer list, and the second-generation truck has now entered service. Aurora validated the Dallas-to-Laredo route only six weeks after beginning supervised autonomous runs there.
Wayve has grown through partnerships and financing. Its latest round brought in $1.2 billion immediately, with further Uber capital tied to milestones. Nissan plans a Tokyo robotaxi pilot, Stellantis intends to integrate Wayve into its supervised-driving platform, and Mercedes-Benz also invested in the company. Those programs have not yet created a large deployed fleet.
Momenta owns the strongest proven growth story. Kodiak is moving fastest in paid driverless activity, while Aurora is expanding the most important operating network.
If you want more recent data on this point, please see our latest autonomous vehicle market report.
Which self-driving startup has won the strongest customer deals?
Momenta has the deepest completed automaker relationships, while Aurora, Kodiak and Nuro have the most consequential driverless customer agreements.
Momenta works with 24 automakers and says it has established relationships with nine of the world’s ten largest vehicle manufacturers. Its disclosed partners include Mercedes-Benz, Toyota, General Motors, SAIC and Ford. More importantly, 68 vehicle models had already reached production.
Aurora’s customer base is smaller but commercially serious. McLane, a Berkshire Hathaway subsidiary, has started using its driverless service. Hirschbach has selected Aurora for a planned fleet of 500 trucks beginning in 2027. Aurora’s own filing describes the Hirschbach agreement as non-binding, so the 500 vehicles remain a customer intention rather than a completed order.
Kodiak’s Atlas relationship carries more present-day weight. Atlas already owns and operates 28 driverless vehicles and has discussed growing the fleet toward 100. Each expansion requires Atlas to put more capital and freight volume behind Kodiak’s software.
Nuro has assembled the largest detailed robotaxi program in the group. Uber, Lucid and Nuro plan to deploy at least 20,000 vehicles over six years. Lucid supplies the Gravity-based vehicle, Nuro provides the driver, and Uber or its fleet partners own and operate the cars. A production-intent vehicle has been unveiled, employee testing has started through the Uber app, and Houston has been named as the second planned market.
Wayve now has several routes to market. Nissan and Uber plan a Tokyo pilot in late 2026. Stellantis has announced a supervised-driving integration targeted for 2028 and is separately exploring global Level 4 robotaxis with Wayve and Uber. Mercedes-Benz is both an investor and development partner.
Waabi’s proposed Uber relationship covers up to 25,000 robotaxis, but there is no named vehicle, launch city or public timetable. Part of Uber’s capital also depends on Waabi reaching development milestones.
This chart, included in our autonomous vehicle market deck, shows how Waymo is winning in autonomous vehicles
Which self-driving startups have real products today?
Momenta, Aurora and Kodiak currently have commercial products; PlusAI and Oxa have genuine working deployments; Wayve, Nuro and Waabi remain closer to launch than to scale.
Momenta’s product is the easiest to recognize as normal automotive software. Automakers pay engineering fees while adapting it to a vehicle and licensing fees once production begins. Consumers then buy the finished car through the automaker’s usual sales network.
Aurora moves customer freight without an onboard driver. The fleet is still small and Aurora’s revenue remains minimal, but the product is performing the task it was designed to perform. Its latest hardware generation is now being installed through a repeatable process at Roush.
Kodiak’s software has become part of Atlas’s daily industrial operation. The trucks work without a driver, carry paying loads and belong to the customer. Kodiak is still developing the safety case required for driverless long-haul highways.
PlusAI has placed a factory-integrated International truck into a live trial operated by Ryder. The truck travels a daily 600-mile route between Laredo and Temple, carrying commercial freight. A safety operator remains present, and fully driverless factory-built trucks are targeted for 2027.
Oxa has completed smaller but authentic deployments. A DHL project at Heathrow covered 1,300 autonomous kilometres in live airport traffic over 14 days. Oxa also works with companies including BP on industrial sites. Small, yes. But real.
Nuro operates nearly 100 Lucid engineering vehicles across several US locations. Select Uber employees have tested the service through the Uber app, and Nuro has received California permits allowing further passenger testing. Public commercial rides have yet to begin.
Wayve’s Tokyo pilot is due later in 2026, while its consumer-vehicle programs start later. Waabi has integrated its driver into Volvo’s autonomous truck and demonstrated driving on highways and surface streets, but it has not disclosed a paid driverless fleet.
Can we trust self-driving performance claims?
Aurora and Kodiak provide the strongest operating proof, but the public numbers do not support a universal safety ranking.
Nearly every company measures something different. Aurora reports public-road driverless miles. Kodiak reports paid driverless hours and loads. Momenta reports production kilometres and success rates for individual manoeuvres. Wayve highlights the number of cities reached. PlusAI publishes an internal safety-readiness percentage. Waabi emphasizes simulated coverage and zero-shot generalization.
Aurora’s driverless miles are especially useful because no person sits behind the wheel ready to take over. The company has also completed a safety case for its second-generation platform, and Edge Case Research recently carried out an outside assessment of Aurora’s safety-assurance approach. Aurora does not publish enough detail about remote assistance, route restrictions and operational pauses to calculate a fully comparable failure rate.
Kodiak’s paid hours show endurance rather than geographic difficulty. Its vehicles operate for long periods in heat, dust and demanding industrial conditions. The controlled environment contains fewer unpredictable road users than a public highway.
Momenta’s production fleet had accumulated 6.85 billion kilometres by the end of 2025. The company reports strong rates for manoeuvres such as lane changes, roundabouts and U-turns, along with active-safety results under C-NCAP and Euro NCAP protocols. Many of these figures come from Momenta or the consultant used in its prospectus.
Wayve says it achieved zero-shot driving in more than 500 cities. Waabi recently reported that its Volvo truck handled a new environment without further training. Both claims are relevant to generalization, but neither shows how often the system fails.
| Startup | Strongest performance proof | What we can reasonably conclude | What remains unknown |
|---|---|---|---|
| Aurora | Nearly 440,000 driverless public-road miles | The software can operate Class 8 trucks without an onboard fallback across several routes | Comparable intervention, delay and availability rates |
| Kodiak | More than 23,500 paid driverless hours | The system can work repeatedly inside a customer’s industrial operation | How well the same stack performs on open highways |
| Momenta | Billions of production kilometres and detailed manoeuvre results | The system has broad exposure and has passed repeated automaker validation | How that record translates into Level 4 driverless safety |
| Nuro | Long driverless history and nearly 100 new engineering vehicles | Nuro has meaningful experience and an active validation fleet | Performance of the new passenger service at commercial scale |
| Wayve | Driving tests in more than 500 cities | The software appears unusually adaptable across locations | Failure frequency under unsupervised commercial operation |
| PlusAI | Daily supervised commercial freight and published readiness metrics | The truck can handle long real-world freight routes | Readiness for operation without a safety driver |
| Waabi | Volvo integration, simulation and zero-shot demonstrations | The AI approach can transfer across selected environments | Independent proof from sustained customer operations |
| Oxa | Live airport and industrial deployments | The software works in defined industrial workflows | Reproducibility across a large number of sites |
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
Which self-driving startup uses capital most efficiently and offers customers the best economics?
Momenta currently has the strongest software economics, while Kodiak has produced the most visible driverless progress with a relatively modest amount of capital.
Momenta’s gross margin reached 71.6% as licensing became a much larger part of revenue. The company earned considerably more money without adding the same amount of delivery cost. Its model also avoids funding a fleet because automakers build the vehicles, dealers sell them and customers own them.
Momenta has raised about $1.8 billion including its IPO and recently generated roughly $335 million in annual revenue. Its latest annual sales equal almost one-fifth of the cumulative capital estimate. That ratio does not measure investor returns, but it shows that customers are already financing part of the company’s growth.
Kodiak’s economics are less transparent, although Atlas’s behaviour is useful evidence. Atlas started with two customer-owned trucks, expanded to 20 and then reached 28. The vehicles have also been working more hours per quarter. Kodiak has used considerably less capital than Aurora and already supports paid driverless work.
Aurora has absorbed more than $3 billion of identifiable private and public capital while producing very little revenue. Its second-generation hardware reportedly costs more than 50% less than the first version and is built to last one million miles. Aurora now has to show that hundreds of revenue-generating trucks can support the spending.
Wayve has closed around $2.5 billion without disclosing meaningful commercial revenue. Nuro has raised more than $2.3 billion and changed strategy after moving away from building its own delivery vehicles. Waabi has closed more than $1 billion despite not yet operating a paid driverless fleet.
Oxa has raised less than most road-autonomy companies and narrowed its focus to industrial work. PlusAI has raised about $720 million and reached supervised commercial freight, though its full driverless launch remains ahead.
Who can scale self-driving software and defend the lead?
Momenta has already proved that it can scale through automotive factories, and its production fleet currently gives it the hardest overall advantage to copy.
Momenta has repeated the production process across 68 models and 24 automakers. Each launch requires vehicle integration, safety testing, cybersecurity work, manufacturing approval and continuing software support. Its vehicles had also generated 6.85 billion kilometres of driving data by the end of 2025, with roughly 630,000 new data clips selected each day.
A competitor can build another driving model, but it cannot immediately recreate that flow of consumer data or Momenta’s completed automaker integrations. Replacing software after a model enters production may also require fresh validation, engineering changes and regulatory work.
Aurora has the strongest scaling plan for fully driverless trucks. Roush now has a formal installation process for the redundant systems and autonomy hardware, with a target production rate of 1,000 trucks a year. Aurora’s advantage also includes custom sensors, terminals, remote assistance, route validation and a formal safety case.
Kodiak has built a repeatable retrofit model with Roush, Bosch and modular SensorPods. Its customer-owned fleet grew by eight vehicles in one quarter. The Atlas relationship has also given Kodiak operational knowledge around maintenance, loading, multi-trailer configurations and remote support.
Nuro has the clearest large robotaxi supply chain. Lucid manufactures the vehicle, Nuro supplies the autonomy, and Uber provides fleet ownership, depots, demand and customer service. Wayve could eventually scale across more brands than any rival through Nissan, Mercedes-Benz and Stellantis.
PlusAI has secured factory programs with International, Scania, MAN, Iveco and Hyundai. Waabi’s Volvo partnership gives it a similarly strong vehicle platform, while its simulator may reduce the physical testing needed for new environments.
AI architecture alone is unlikely to remain a lasting moat because the leading companies are adopting similar end-to-end learning, simulation and data-selection methods. Production data, customer workflows, safety history and switching costs should prove harder to reproduce.
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
Which self-driving startup has the strongest momentum lately?
Aurora has the strongest operational momentum right now, with Kodiak close behind and Nuro making the most concrete progress toward a new robotaxi launch.
Aurora has moved beyond adding miles to upgrading the product itself. Its second-generation driverless trucks are now operating without a person in the cab, the network covers ten routes, and Roush is preparing for a much higher installation rate. Customer demand also looks broader than it did a year ago.
Kodiak’s latest figures show that its industrial deployment is accelerating rather than settling into a permanent pilot. As seen above, the customer-owned fleet increased to 28 vehicles while paid hours more than doubled in one quarter. Kodiak also launched freight work with Roehl Transport and announced a military-vehicle collaboration with General Dynamics Land Systems.
Nuro has lately turned its robotaxi plan into something more tangible. The partners have unveiled a production-intent vehicle, built a fleet of nearly 100 engineering cars, started Uber-app testing with employees, obtained additional California permission and identified Houston as a second market.
Wayve continues to collect strategically valuable programs. Its Nissan-Uber pilot is expected to begin in Tokyo before the end of 2026, while Stellantis has added both a supervised-driving project and a separate robotaxi collaboration.
PlusAI has started a daily 600-mile commercial freight trial. Waabi recently demonstrated its Volvo truck on highways and surface streets. Oxa raised another $103 million and sharpened its focus around industrial autonomy.
Momenta’s strongest recent development is financial: its listing provided fresh capital and revealed how quickly the licensing business had grown.
Which self-driving software startups are actually ahead?
Momenta is ahead overall today, Aurora is the closest challenger, and Kodiak has earned a place in the leading group.
Momenta takes first place because it has built the largest deployed product, the largest revenue base, the strongest licensing economics and the widest completed automaker footprint. Most of its software remains supervised, which prevents it from winning the narrower Level 4 comparison.
Aurora ranks second because it leads the technically harder driverless race. Its trucks operate on public highways without anyone behind the wheel, and the company now has a second-generation platform designed for larger production. Aurora would overtake Momenta if it converted that lead into hundreds of active trucks and substantial recurring revenue.
Kodiak ranks third after its latest expansion. Customer ownership, paid work and rising vehicle utilization give it stronger commercial proof than the more heavily funded companies below it. Its industrial routes are easier than public highways, so moving long-haul operations beyond supervised testing would strengthen its case considerably.
Wayve holds fourth place. Its technology has attracted the strongest group of global automakers among the private challengers, and its cross-country testing is unusually broad. The gap with the top three comes from deployment.
Nuro ranks fifth. The Lucid-Uber program now has a production-intent vehicle, permits, a large test fleet and named launch markets. Its earlier delivery operations also give it more driverless experience than most private rivals.
PlusAI ranks sixth because its software is carrying commercial freight inside a factory-integrated truck, and its relationships span several major truck manufacturers. Waabi follows in seventh place: its technology, simulator, Volvo relationship and financial backing make it a serious challenger, but its disclosed commercial operation remains limited.
Oxa ranks eighth overall but occupies a sensible niche. Industrial autonomy may commercialize earlier than open-road autonomy because routes are repetitive, labour shortages are immediate and regulation is simpler. Oxa still needs larger and more repeatable deployments.
Momenta’s more than 733,000 installations place its current business at a completely different scale from the rest of the field. Aurora’s nearly 440,000 public-road driverless miles remain the strongest evidence that a startup-built driver can operate in a difficult environment without human supervision. Kodiak lacks Momenta’s volume and Aurora’s road complexity, but real customers already own its vehicles and put them to work.
| Rank | Startup | Why it ranks here |
|---|---|---|
| 1 | Momenta | The strongest combination of production scale, revenue, licensing growth, margins and completed automaker programs |
| 2 | Aurora | The leading public-road Level 4 platform, now supported by second-generation trucks and a growing route network |
| 3 | Kodiak | The best proof that customers can own driverless vehicles and use them for continuous paid work |
| 4 | Wayve | The strongest cross-OEM platform challenger, held back by limited commercial deployment today |
| 5 | Nuro | A credible, increasingly concrete robotaxi program backed by Lucid and Uber, following a costly strategic pivot |
| 6 | PlusAI | Real supervised freight, broad truck-manufacturer support and a defined path toward factory-built driverless vehicles |
| 7 | Waabi | Strong AI, simulation, funding and partners, but commercial proof still trails the size of its ambitions |
| 8 | Oxa | A credible industrial-autonomy specialist whose projects have not yet reached broad, repeatable scale |
If you want more recent data on this point, please see our latest autonomous vehicle market report.
This chart, featured in our autonomous vehicle market deck, shows the share of revenue generated by each customer segment in the autonomous vehicle market
OUR METHODOLOGY
This analysis compares eight companies whose core product is a self-driving software platform that can be licensed, installed or operated inside vehicles made, owned or used by other businesses: Momenta, Aurora, Kodiak, Wayve, Nuro, PlusAI, Waabi and Oxa.
We excluded companies whose autonomy business sits inside a much larger technology or automotive group, including Waymo, Zoox, Tesla, Mobileye, Nvidia and Cruise. We also excluded transport-service operators such as May Mobility and Gatik when the value and maturity of the underlying software driver could not be separated cleanly from the fleet or logistics service.
We gave the most weight to completed production installations, customer-owned vehicles, paid driverless operations, public-road driverless mileage, reported revenue, licensing growth and repeat deployments. Announced partnerships, future fleet targets and milestone-based financing were treated as evidence of a route to market, but not as equivalent to vehicles already working for customers.
The companies report different operating measures, so we did not force them into one mechanical score. Aurora’s driverless public-road miles show Level 4 capability in a difficult environment; Kodiak’s paid hours and loads show sustained customer use; Momenta’s production installations and licensing revenue show automotive software scale; and factory integration or customer ownership shows whether a program can move beyond a vendor-run pilot.
Capital figures combine identifiable private rounds, public financing and IPO proceeds where relevant. Milestone-based commitments are shown separately from capital already closed because a proposed investment is not the same as cash received. The figures remain approximate since companies classify strategic investments, PIPEs, grants and other financing differently.
Safety and performance claims were assessed according to what they demonstrate directly. We did not treat city counts, simulated coverage, production kilometres, driverless road miles or paid industrial hours as interchangeable, and we did not use company-reported figures to create a universal safety ranking that the available data cannot support.
The final ranking was based on the maturity, commercial value and repeatability of the evidence rather than the number of sections won. Present-day deployment and customer use carried more weight than future ambition, while software revenue, factory integration, operating scale and a credible path to broader rollout separated the leaders from well-funded challengers.
Key sources include Momenta’s Hong Kong listing announcement and official Chinese-language listing materials; Aurora’s second-generation truck update, first-quarter 2026 shareholder letter and 2025 results; Kodiak’s first-quarter 2026 results and full-year 2025 results; Wayve’s Series D announcement; Nuro’s Lucid and Uber robotaxi program; PlusAI’s SEC-hosted presentation; Waabi’s funding and Uber announcement; Oxa’s Series D update and DHL deployment report; and SAE International’s driving-automation taxonomy.
This chart, included in our autonomous vehicle market deck, shows how robotaxi platform technology has evolved over time
Related blog posts
- Robotaxis: which startup is ahead?
