Robotaxi software platforms take shape -

Robotaxi software platforms take shape


Robotaxi software platforms are making progress, but so far none have generated meaningful revenues.

We review how the leading robotaxi software platforms are faring in terms of partners and testing activities  The format used here is similar to a recent post on autonomous truck software

At least four factors help determine the future success of robotaxi software platforms: key investors, auto OEMs, robotaxi partners and city tests and trials. Ultimately, success will be measured in revenues, which remain elusive.

The chart below summarizes 11 robotaxi software platforms, including information for each of the four key factors organized as one row of data for each category. (There was not room to include other robotaxi software platforms such as DeepRoute, Didi, Yandex, Zoox and others.)

click for full size image

Robotaxi software platforms are listed first in red; hardware platforms in blue.

Key investors primarily include auto and transportation-related companies. Leading venture capitalists are very important but are not listed. Among the companies listed, Aurora Innovation went public in 2021.

Customer relationships with car OEMs are listed in the third row in blue. It is likely that many more OEM relationships will be announced or will be added soon. Some auto OEMs are likely to use more than one robotaxi platform—most likely a different partner in China.

Relationships with robotaxi operators appear in green in the fourth row. Listed primarily are software platform developers, plus ride-hailing companies. Many OEMs are expected to become future robotaxi operators.

The last row list cities where the robotaxi software is being tested or where robotaxi services are underway. Those activities are likely to grow quickly over the next years.


Waymo remains the leader in AV software platform development given its first-mover advantage in 2009. It became an independent subsidiary of Google parent Alphabet in 2017. Waymo has received around $5.75 billion in funding from investors.

The Waymo Driver software platform is focused on three AV use cases: robotaxis, goods deliveries and autonomous trucks.

The robotaxi service started in Phoenix area in December 2018 with safety drivers. Waymo launched driverless services for a small portion of its rides in September 2019, with nearly all becoming driverless in October 2020. Waymo’s expansion in Phoenix has been slower than expected with long wait times reported recently.

Waymo initiated trials in San Francisco in August 2021. Several hundred passengers are booked, with a waiting list in the thousands. Robotaxi service is likely to commence in 2022. Waymo has begun mapping and testing in New York City and Los Angeles.

OEM partners include Nissan, Renault, Stellantis and Volvo. Waymo and Chrysler have been partners for several years.

Waymo is the leader in robotaxi-related testing in California with 4.8 million miles driven between 2015 and 2020. It’s also the leader in AV disengagement data, or average miles driven before a safety driver takes over. Waymo reached nearly 30,000 miles between disengagements in California in 2020.

Aurora Innovation

Aurora Driver focuses on robotaxis and autonomous trucks. The company was founded in 2017 by three pioneers in the AV industry. We reviewed its special-purpose acquisition company (SPAC) IPO here.

Toyota and Uber are development partners and investors. Toyota and Aurora have committed to a long-term partnership. Aurora also has a 10-year agreement to using Uber data from its ride-hailing operation, which was included as part of Aurora’s acquisition of Uber’s AV development group in December 2020.

Aurora has relied mostly on virtual testing before launching public road testing. Its robotaxi road miles are primarily in Mountain View, San Francisco and Pittsburgh. It plans robotaxi road testing in Dallas.

It also expects to generate revenue from its mobility-as-a-service (MaaS) technology beginning in 2024, a year later than anticipated autonomous truck revenue.


San Francisco-based Cruise was founded in 2013. It was acquired by GM in February 2016 for about $1 billion. Cruise has received around $10 billion in funding from GM, Honda, Microsoft, Softbank, Walmart and other investors. It has about 1,900 employees.

According to the website TechCrunch, Cruise is designing two custom chips as part of it AV platform, one for sensor data processing and the other for machine learning algorithms.

It’s two major OEM customers are GM and Honda. Cruise has been testing robotaxis in Phoenix as well as San Francisco, the latter since 2016. An office in Seattle indicates future testing there.

The company has been testing a food-delivery service during the pandemic, providing more than 1 million meals to food banks. Walmart and Cruise started last-mile groceries delivery tests in Scottsdale, Ariz., in late 2020.

Last March, Cruise acquired Voyage, a startup focused on fixed-route MaaS use cases, indicating Cruise has future plans in mobility services.

During CES 2022, CEO Mary Barra said GM would introduce consumer AVs in the 2024-25 timeframe. GM also provided private showings at CES of a self-driving luxury concept car. It included the goal of 1 million self-driving cars by 2030 in its October 2021 investor presentation.

Through 2020, Cruise testing totaled 2.2 million miles in San Francisco, mostly by robotaxis and some goods-delivery miles. By now Cruise should be over 3 million miles.

It had accumulated more than 28,500 miles between AV disengagements in 2020, according to California Department of Motor Vehicles data.

After five years of testing in San Francisco, Cruise started driverless AV trials in December 2020. It appears 2022 will be the year driverless robotaxi trials for Cruise begin in San Francisco.

Pittsburgh-based Argo was founded in 2016. It has so far attracted about $3.6 billion from its investors, including Ford, Lyft and VW. Argo also offers its AV software platform to other customers—including other auto manufacturers. Argo is considering an IPO as early as this year.

It is focused on two AV use cases: robotaxi and goods delivery. Ford is expected to use Argo for future robotaxis and goods AVs. VW is likely to have a similar strategy, initially in Germany.

Argo has so far released less information than its competitors.

What is known is that the company used its AV test vehicles to deliver thousands of Covid-related goods in Pittsburgh.

It has also been testing in at least four other cities: Austin, Miami, Munich and Washington, DC. Argo says it will partner with Lyft to deploy robotaxis in Miami this year with safety drivers.


Mobileye was founded in 1999 with headquarter in Jerusalem, Israel. It went public in 2014 with a valuation of $5.3 billion. In the last decade Mobileye has been the clear leader in ADAS products based on its EyeQ SoC product family. Mobileye’s cumulative EyeQ sales have topped 100 million units, including 28 million in 2021.

Intel acquired Mobileye in 2017 for more than $15 billion. The chip maker plans to spin off Mobileye in an IPO this year, with Intel retaining majority ownership.

Mobileye is leveraging its leading ADAS status for future L3 and L4 AVs. It has two L3 design wins: Honda Legend in Japan and BMW 7-series coming in 2022. It is likely that other existing ADAS customers will use Mobileye L3 and L4 products.

Mobileye has a unique strategy for L4 AVs: robotaxis for passengers and goods delivery starting in 2022 and consumer L4 AVs beginning in 2024. It has multiple design wins for goods and people robotaxis, including Sixt, RATP Group, Transdev, Udelv, Willers and Marubeni.

Mobileye has one publicly disclosed design win for a consumer L4 AV—Zeeker, (a brand from China’s Geely Auto Group)—with a production starting in 2024. It also has an unnamed design win for a consumer L4 AV.

In April 2021, the company announced commercial availability of its Mobileye Drive, self-driving system for L4 MaaS use cases. The platform is based on three core technologies—Road Experience Management, Responsibility-Sensitive Safety (RSS) and two independent-perception subsystems called True Redundancy—a camera- and lidar-radar based system.

During 2022 CES, Mobileye introduced its EyeQ Ultra SoC that implements its RSS safety algorithms for L4 systems.

Mobileye is currently testing in Detroit, Jerusalem, Munich, New York, Paris and Tokyo.

The company’s CES 2022 presentation, including details on its ADAS and AV technologies, design wins and company strategy is here.


Motional is a $4-billion joint venture between Aptiv and Hyundai formed in early 2020—with each owning 50 percent. Motional is based primarily on Aptiv’s AV technology. Aptiv moved early to acquire AV startups such as Ottomatika in 2013 and nuTonomy in 2017. It leveraged those acquisitions into a strong AV software platform for robotaxi deployments.

Lyft has been the key partner for Motional in Las Vegas. It is likely that other operators will use Motional robotaxi software platform.

Nevada granted Motional permission to use driverless AVs in November 2020. It is likely that Motional will use the permit for initial testing in Las Vegas.

In December 2020, Motional said it would deploy driverless AVs on Lyft’s ride-hailing network starting in 2023 in multiple markets.

Motional operates more than 100 AVs in multiple cities, including Las Vegas, Boston and Singapore. By January 2020, its AVs completed more than 100,000 robotaxi trips with safety drivers aboard. With Hyundai’s 50-percent ownership stake, it likely means robotaxi testing and deployment in South Korea.

Chinese robotaxi software platforms

Here we update an earlier post covering five Chinese companies developing robotaxi software platforms. Development, testing and deployment in China includes:

  • AV permit by city, only for a specific area, expanding over time. More than 70 companies have received AV testing licenses in one or more Chinese cities.
  • Usually fixed pickup and drop-off points. This simplifies AV driving by establishing fixed routes that can avoid the most difficult traffic situations.
  • A requirement for safety drivers as backups.
  • Commercial tests with customer fees as the next step in multiple cities.
  • Driverless testing has emerged recently in a few cities. Teleoperation backup is required.
  • Plans for low-volume production of robotaxis with Chinese OEMs, now underway.
  • Planning for volume deployments with thousands of robotaxis per city.

Not all Chinese software developers are included here. Didi, the leading ride-hailing company, has its own robotaxi development. Several robotaxi startups such as DeepRoute and G-PAL are worth mentioning. Many OEMs in China are also investing in AV testing and often work with multiple AV software platform companies.

AutoX is included though its headquarters is in California. Testing and deployment remains in China.

Apollo Go

Apollo is part of Baidu, China’s largest internet search company. Apollo is China’s leading AV developer. It has built an open source ecosystem, with over 210 companies participating. It includes both domestic and foreign auto OEMs, Tier-1s and much of the automotive supply chain.

Apollo offers robotaxi services in Beijing, Cangzhou, Changsha, Guangzhou and Shanghai. It has tested AVs in 27 cities, including California. Apollo plans to expand its Chinese robotaxi service to 65 cities in 2025 and 100 cities by 2030.

The company is working with at least five auto OEMs to produce AVs: BAIC, FAW, Geely, Great Wall and Volvo.

The Apollo Moon robotaxi is based on a battery-EV design derived from BAIC’s premium brand, Arcfox. The Moon robotaxi is equipped with 13 cameras, five radars and two lidars, with computing power reaching 800 TOPS. It also features redundant hardware for safety. Driverless Apollo Moon vehicles began operating in Beijing last October, offering public robotaxi services.

Baidu has a joint venture with Geely. The partners plan to announce their first L4 robotaxi in the first half of 2022, with production and deployment starting in 2023.

Apollo’s cumulative L4 road miles were expected to top 13 million by the end of 2021 based on past growth trends. Virtual test miles in the past have been about 100 times higher than road miles.

Apollo had received 411 AV test permits in China as of the end of September, or nearly half of all AV permits in China.


San Jose-based AutoX was founded in 2016, so far attracting $160 million in venture funding.

AutoX has partnered with Chrysler (now part of Stellantis) to develop robotaxis. The partners will launch a fleet in China, using the Chrysler Pacifica and AutoX AV platform. A dedicated production line for the Pacifica was launched last July in China.

In December, AutoX announced that L4 robotaxis based on the Chrysler Pacifica are in production. AutoX plans to deploy thousands of robotaxis for its China service.

It is also working with two Chinese OEMs, BYD and Dongfeng.

Trials are underway in Beijing, Guangzhou, Shanghai, Shenzhen, with plans for more testing in Wuhu and Wuhan. AutoX claims to operate China’s largest robotaxi service area, covering 65 square miles in Shenzhen, where testing began in early 2021.

AutoX has also tested in California since 2018. It achieved among the best results in California’s AV testing in terms of mileage between disengagements during 2019 and 2020.


Momenta sells ADAS and semi-automated driving software to auto OEMs while investing in L4 autonomous vehicle software. It is developing driver monitoring systems, including products and services for fleet operators.

Momenta was founded in 2016 with headquarters in Beijing. It also has an office in Stuttgart, Germany. Funding exceeds $2 billion, which tops other Chinese AV startups. Auto company investors include Bosch, GM, Mercedes-Benz, SAIC and Toyota. Tencent, a leading Chinese internet company, is also a major backer.

Momenta claims many Chinese and foreign OEM customers for its ADAS products, including BYD and SAIC.

Momenta Go started trial operations with safety drivers in October 2020 in Suzhou. Momenta will test driverless robotaxis in 2022.

Momenta has an impressive list of auto-related investors. The combination of ADAS and AV development is likely the reason since ADAS deployment generates loads of data for AV and map development, testing and deployment. Momenta is applying a strategy similar to Mobileye’s—a good move. was founded in 2016, with headquarters in Guangzhou, China, and Fremont, Calif. Investment has surpassed $1.1 billion. It canceled a SPAC IPO in August in response to restrictions imposed by the Chinese government.

FAW and Toyota are among its nearly 30 investors. FAW, GAC and Toyota are planning to use AV software from, which began testing GAC-based L4 vehicles in September 2019. and Toyota launched AV pilot testing in August 2019 in China. The following July, the company began testing in Shanghai with a fleet of Lexus RX vehicles equipped with AV technology.

It has also fielded more than 200 AVs for testing and robotaxi services in Guangzhou, Shanghai and Beijing. Testing with passengers in Guangzhou started in late 2018. The Chinese testing effort surpassed 4 million miles in 2021. received permission for driverless testing in Beijing in October 2021, and paid robotaxi services in Beijing began in November. is testing in California where it has the largest AV effort of any Chinese developer. By end of 2020, it had accumulated 225,000 miles of testing data via 29 AVs operating in Irvine, Milpitas and Fremont. It obtained a permit in May 2021 to test six AVs without a safety driver on designated roads in the three Californian cities. However, testing was suspended in December 2021 due to a crash with a center divider in Fremont.


Founded in 2017 with headquarters in Guangzhou, China. Venture funding to date totals about $620 million. Investors include two auto manufacturers, Yutong and the Renault-Nissan-Mitsubishi Alliance.

Renault-Nissan-Mitsubishi became an investor in October 2018, with additional funding in June 2021. WeRide received a strategic investment of $200 million in December 2020 from Yutong, a leading Chinese bus manufacturer.

OEM partners include Dongfeng, GAC, Jiangling Motors, Nissan and Yutong. WeRide use the Dongfeng Fengshen E70 for AV testing and robotaxi services.

Last November, WeRide announced strategic cooperation with GAC Group and its Ontime subsidiary mobility service platform to expand robotaxi commercialization.

WeRide’s strategy includes multiple AV use cases: robotaxis, last-mile goods delivery and autonomous trucks, with testing in all three segments. Robotaxis constitute the core strategy. WeRide has over the past year developed a robo-van and a mini-robobus with partners.

WeRide began AV testing in November 2018 in Guangzhou, expanding to a public robotaxi service in November 2019 that covers 144 square kilometers. The pilot program is a joint venture with a large Chinese taxi operator, with over 10,000 taxis.

WeRide launched driverless testing in Guangzhou in July 2020. Driverless AVs include teleoperation by a remote operator. WeRide’s fleet includes at least 10 driverless AVs. WeRide has more than 120 AVs dedicated to robotaxi operations and AV testing.

Bottom line

Robotaxi software development is making progress. Many trials with passengers are underway across the U.S. and China, some with paying customers.

Driverless trials have also begun, with Waymo still leading in Phoenix. The coming year will see considerable expansion in robotaxi trials by city, region and technology.

Still, robotaxi software platform developers have yet to generate meaningful revenue.

>> This article was originally published on our sister site, EE Times.

Egil Juliussen has over 35 years’ experience in the high-tech and automotive industries. Most recently he was director of research at the automotive technology group of IHS Markit. His latest research was focused on autonomous vehicles and mobility-as-a-service. He was co-founder of Telematics Research Group, which was acquired by iSuppli (IHS acquired iSuppli in 2010); before that he co-founded Future Computing and Computer Industry Almanac. Previously, Dr. Juliussen was with Texas Instruments where he was a strategic and product planner for microprocessors and PCs. He is the author of over 700 papers, reports and conference presentations. He received B.S., M.S., and Ph.D. degrees in electrical engineering from Purdue University, and is a member of SAE and IEEE.

Related Contents:

For more Embedded, subscribe to Embedded’s weekly email newsletter.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.