Hubbry Logo
Argo AIArgo AIMain
Open search
Argo AI
Community hub
Argo AI
logo
8 pages, 0 posts
0 subscribers
Be the first to start a discussion here.
Be the first to start a discussion here.
Argo AI
Argo AI
Argo AI
View on Wikipedia
from Wikipedia

Argo AI LLC was an autonomous driving technology company headquartered in Pittsburgh, Pennsylvania.[2][3] The company was co-founded in 2016 by Bryan Salesky and Peter Rander, veterans of the Google and Uber automated driving programs.[4] Argo AI was an independent company that built software, hardware, maps, and cloud-support infrastructure to power self-driving vehicles. Argo was mostly backed by Ford Motor Co. (2017)[5] and the Volkswagen Group (2020).[6] At its peak, the company was valued at $7 billion.[7]

Key Information

In October 2022 it was announced by Ford that the company would be disbanded and some employees would be split between VW and Ford. Argo’s technology will be salvaged and further developed in-house by Ford and VW.[8][9][10] Ford stated their intent to change the focus of development from Level 4 autonomous driving to Level 3 and Level 2+.[11][12]

History

[edit]

Founding

[edit]

Argo AI was co-founded in November 2016 by roboticists Bryan Salesky, CEO of the company, and company president Peter Rander. As of July 2020, the company employed over 1000 employees with offices[13] in Pittsburgh, Detroit, Palo Alto, Cranbury, NJ[14] and Munich, Germany.[15] In June 2020, Argo was valued at $7.25 billion.[16]

Salesky was born in Woodhaven, MI outside of Detroit and settled in Pittsburgh, Pennsylvania, where Argo is based.[4] He earned a Bachelor of Science in Engineering from the University of Pittsburgh in 2002 and began working at Union Switch & Signal, a supplier of railroad signaling equipment, where he worked on anti-collision software for trains.[4] From 2004 to 2011 he worked at the National Robotics Engineering Center (NREC) at Carnegie Mellon Robotics Academy, where Rander was his boss.[4]

In 2007, Salesky and Chris Urmson[4] led Carnegie Mellon’s software development team for the third DARPA driverless car competition dubbed the Urban Challenge.[17] From 2013 - 2016, Salesky served as director of hardware development for Google’s autonomous vehicle initiative, now known as Waymo.[17][4]

Rander, a Michigan native [18] and University of Detroit Mercy alumnus [19] earned his MS and Ph.D. degrees in Robotics at Carnegie Mellon University (CMU) and worked alongside Salesky at the National Robotics Engineering Center.[20] From 2015 to 2016,[21] Rander was the engineering lead for Uber’s Advanced Technology Group (ATG); Uber’s self-driving car unit.[22]

Argo AI was initially funded in late 2016 through a small seed round from an undisclosed source.[15] In February 2017, Ford Motor Company announced that it was to invest $1 billion in Argo AI over the next five years, to develop a virtual driver system for the automaker's autonomous vehicle coming in 2021.[5] Ford became the company’s largest shareholder, enabling them to hire 200 additional employees including members of Ford's R&D team.[23]

Volkswagen investments and disbandment

[edit]

In June 2020, Volkswagen invested $2.6 billion into Argo AI. Under the terms of the deal, Volkswagen committed $1 billion in cash to Argo AI and its Munich-based Autonomous Intelligent Driving (AID) unit - valued at $1.6 billion.[24] The AID technology unit, consisting of over 200 employees, was first launched to develop autonomous vehicle technology for VW automobiles. According to the terms of the deal, VW was to purchase Argo AI shares from Ford for $500 million over three years, while Ford would invest the remaining $600 million of its $1 billion cash commitment to the company. In executing the agreement, the Munich-based AID team was integrated into Argo AI to form the company’s European engineering center.[24]

Between the partnerships, Argo raised over $3.6 billion.[15] Argo’s ownership was shared among its largest shareholders, Ford and Volkswagen, each of whom owned 42 percent of the company.[25] In 2017, its board of directors was composed of two Ford seats, two Volkswagen seats, and three Argo seats.[24]

In July 2021, Lyft invested in Argo with a plan to launch services on Lyft’s ride-hailing network in Austin and Miami. The partnership with Lyft boosted Argo’s valuation to $12.4 billion.[26]

In March 2022, Argo AI announced the opening of a new engineering and development office in Los Angeles.[27]

In October 2022, after failing to secure further investment from Amazon,[28] Ford announced that the company would be disbanded after posting an $827 million net-loss for the third quarter. Ford assumed a $2.7 billion accounting charge and wrote off a cash investment of about $500 million for Argo AI.[8] Similarly, Volkswagen expensed €1.9 billion [29]

A number of resources from Argo AI were transitioned to Ford's ADAS effort Latitude AI, and VW's ADAS efforts CARIAD.[30] VW also continued it autonomy efforts with its ID BUZZ platform with Mobileye, an alternative SDS supplier.[31]

Technology

[edit]
A Ford Escape with Argo AI hardware

Argo claimed its technology would deliver Level 4-capable Self-Driving Systems (SDS) for autonomous driving (AV) vehicles.[32] Autonomous driving categories are defined by the Society of Automotive Engineers and have been adopted as industry standards by the U.S. Department of Transportation.[33]

The technology developed by Argo comprises the entire self-driving system, including the software and hardware computing platforms, sensors, cameras, radar, and light detection and ranging radar (LIDAR).[15] In October 2017, Argo purchased Princeton Lightwave, a producer of lidar technology, based in Cranbury, New Jersey.[34] In 2019, Argo introduced its third-generation test vehicle, based on the Ford Fusion Hybrid, which incorporates higher-resolution cameras with wider dynamic ranges, increased computer processing power, and heating and cooling systems improvements.[23]

Argo planned to deliver autonomous driving technology for shared fleets rather than personal ownership with applications including ride-hailing and goods delivery services.[35] 

Road testing

[edit]

Argo has self-driving vehicle testing locations in Pittsburgh and Dearborn, Michigan as part of its research and development initiatives. In early 2018, Argo began testing vehicles in Miami, Florida marking the first time vehicles were tested outside the company’s home territories of Detroit and Pittsburgh.[36] Additional test sites for the Argo and Ford collaboration included Washington, D.C.,[37] and Austin, Texas.[38] Argo began testing out of its Palo Alto engineering center after receiving its permit from the California Department of Motor Vehicles in January 2019.[39] The expansion to additional urban test sites allowed the company to study the unique driving behavior and road culture of each city.[40] Argo’s testing included millions of simulations to represent various weather, infrastructure, pedestrian, and other conditions.[35]

By 2021 Argo AI was testing the VW IDBUZZ on public roads in Munich. [41] In May 2022, Argo AI started testing on public roads in Austin and Miami using a modified Ford Escape Hybrid.[42][43]

University research

[edit]

In 2017, Argo announced university partnerships with Carnegie Mellon University and the Georgia Institute of Technology to research computer vision and machine learning technologies.[44]

In June 2019, Argo announced a $15 million investment over five years to create the Carnegie Mellon University Argo AI Center for Autonomous Vehicle Research, which would have focused on advanced algorithms for autonomous vehicles, specifically for advanced perception and decision-making capabilities.[45][20] The announcement follows the company’s introduction of its open source data set, a set of highly curated maps and data released to further the study of autonomous vehicle research and development.[46] As of 2020, research topics slated for study at the research center included smart sensor fusion, 3D scene understanding, urban simulation, map-based perception, behavior prediction, and software validation.[22]

See also

[edit]

Aurora Innovation

References

[edit]
Revisions and contributorsEdit on WikipediaRead on Wikipedia

Argo AI

View on Grokipedia
from Grokipedia
Argo AI was an American autonomous vehicle technology company founded in 2016 by Bryan Salesky, formerly of Uber's Advanced Technologies Group and Google's project, and Peter Rander, previously with Ford's research division. Headquartered in , with additional engineering centers in ; ; and Munich, Germany, the company developed a full-stack autonomous driving platform including software, sensors, high-definition mapping, and simulation tools aimed at achieving Level 4 autonomy for commercial applications such as ride-hailing and delivery services. The company launched with a landmark $1 billion investment from Ford Motor Company in 2017, which took a majority stake and integrated Argo's technology into its vehicle development pipeline. In 2019, Volkswagen Group added $2.6 billion in funding, expanding Argo AI's operations to Europe and valuing the startup at approximately $7.25 billion at the time. Overall, Argo AI raised more than $3.6 billion from its primary automotive backers, enabling rapid scaling to over 2,000 employees by 2022. Key partnerships included collaborations with Lyft for ride-hailing integration, while the company conducted extensive testing on Ford and Volkswagen vehicles across more than seven cities, including Miami, Austin, Washington D.C., and Hamburg. Argo AI's technology stack emphasized scalable perception, planning, and control systems using and lidar-based sensing, with notable achievements including the deployment of fully driverless vehicles for public rides in select U.S. markets and preparations for Volkswagen's MOIA ride-pooling service in by 2024. However, despite these advances, the company faced escalating costs, with Ford recording a $2.7 billion impairment on its investment amid regulatory hurdles in achieving widespread commercialization. In October 2022, Ford and announced the shutdown of Argo AI, citing a pivot toward more achievable Level 2 and Level 3 autonomy features amid economic pressures and slower-than-expected progress toward full self-driving. Ford absorbed about 400 employees and key for its BlueCruise hands-free highway driving system, while integrated around 220 staff and technology into its Cariad software unit for ID. Buzz autonomous shuttles. The closure marked a significant setback for the autonomous vehicle sector but underscored the high barriers to scaling services. Following the wind-down, Salesky and Rander founded Stack AV in 2023, a new Pittsburgh-based startup focused on autonomous trucking backed by SoftBank.

History

Founding and early development

Argo AI was founded in 2016 as an independent company focused on autonomous vehicle technology, receiving primary backing from through a $1 billion investment announced in 2017 that gave Ford a majority stake. The company received an initial $1 billion investment from Ford, to be deployed over five years, to accelerate development of self-driving capabilities. This funding supported the creation of a dedicated team to build software and hardware for SAE Level 4 autonomous systems, aiming for commercial deployment by 2021. The company was co-founded by CEO Bryan Salesky, a former director of hardware development on Google's project and a veteran of Uber's Advanced Technologies Group, and Peter Rander, who served as and previously led Uber's autonomous vehicle engineering efforts after working on Google's self-driving team. Both founders, alumni of Carnegie Mellon University's National Robotics Engineering Center, shaped Argo AI's vision around scalable Level 4 that could operate without human intervention in defined urban environments. Their combined expertise from leading robotics and AI initiatives at major tech firms positioned the startup to integrate automotive with advanced . Headquarters were established in , , capitalizing on the city's burgeoning autonomous vehicle ecosystem and close ties to , a hub for robotics research. This location facilitated recruitment of talent from the region's academic and tech communities. By the end of 2017, Argo AI had grown its initial team to over 200 employees across offices in , southeastern , and the Bay Area. Early efforts centered on internal simulations and virtual testing environments to validate algorithms before real-world deployment. The company prioritized a modular autonomous vehicle stack designed for compatibility across multiple (OEM) platforms, enabling broader adoption beyond Ford. Later investments from partners like in 2020 further expanded this multi-OEM approach.

Investments and expansion

Following its initial $1 billion investment in , Ford committed an additional $600 million in 2019 as part of its ongoing funding pledge to Argo AI, with further investments in 2020 including a preferred security stake to support the company's growth. In June 2020, completed a $2.6 billion investment in Argo AI—comprising $1 billion in cash and the contribution of its Munich-based Autonomous Intelligent Driving unit, valued at $1.6 billion—establishing a 50/50 ownership partnership with Ford and valuing the company at approximately $7.25 billion. This infusion enabled Argo AI to accelerate its development of autonomous driving technology for both partners. The funding fueled significant operational expansion, with Argo AI growing its workforce to over 2,000 employees by 2022. The company opened additional offices in Detroit, Michigan, to support collaboration with Ford; , to tap into talent; ; and Munich, Germany, following the integration of Volkswagen's autonomous driving team. Amid this growth, Argo AI shifted strategically toward commercialization, prioritizing services through partnerships like the 2021 collaboration with to deploy autonomous vehicles in cities such as and Austin. The company also focused on integrating its technology into original equipment manufacturer (OEM) vehicles, including plans to enable fully autonomous operations in Ford's Transit vans for commercial applications like last-mile delivery. However, these ambitions were tempered by escalating financial pressures, including substantial expenses that highlighted the challenges of scaling autonomous vehicle technology.

Shutdown

On October 26, 2022, Ford Motor Company and Volkswagen Group announced the shutdown of Argo AI, the autonomous vehicle startup they had jointly backed since 2017, citing a strategic shift away from developing fully autonomous robotaxi technology toward more near-term advanced driver-assistance systems (ADAS). The decision was driven by escalating development costs, slower-than-expected progress toward commercialization, and Argo AI's inability to secure additional investors beyond its primary backers, Ford and Volkswagen, who had collectively invested over $3.6 billion. Ford CEO Jim Farley emphasized that the company would redirect resources to Level 2 and Level 3 ADAS features, which are seen as more feasible in the current market, while Volkswagen similarly pivoted to in-house efforts through its software subsidiary Cariad. The closure impacted approximately 2,000 employees across Argo AI's operations in the United States and , with Ford absorbing about 400 employees and key , while integrated around 220 staff into its Cariad software unit. Those not retained received a comprehensive , including an annual performance award, a transaction bonus upon deal closure, continued , and additional termination benefits, described by employees as generous. While specific relocation support was not detailed publicly, the absorption process allowed for transitions to Ford's ADAS division or Volkswagen's Cariad unit, potentially involving moves to facilities in , or , . Argo AI's technological assets were divided between its parent companies for integration into ongoing projects. Ford planned to incorporate Argo's high-definition mapping and simulation technologies to enhance its BlueCruise hands-free highway driving system, accelerating development of Level 2+ features for future vehicles. , meanwhile, absorbed select and motion planning software into Cariad, along with around 200 employees from Argo's office, to support highly automated driving functions in partnership with suppliers like Bosch. The wind-down process was expected to conclude by the end of 2022, with no ongoing operations for Argo AI as an independent entity. The shutdown triggered significant financial write-downs for both automakers in late 2022. Ford recorded a $2.7 billion non-cash, pretax impairment charge on its investment in Argo AI, contributing to an $827 million net loss for the third quarter. recognized a €1.9 billion impairment loss on its equity stake, impacting its other financial results and reflecting the full write-off of the . This event occurred amid broader industry challenges, including regulatory delays for autonomous vehicle deployment, an economic slowdown reducing investor appetite for high-risk AV projects, and intensifying competition from established players like Tesla's Full Self-Driving and Alphabet's , which had advanced further in commercial services.

Technology

Autonomous driving platform

Argo AI's autonomous driving platform, referred to as the Argo AV stack, features a modular tailored for Level 4 in complex urban settings. This stack integrates four core modules: , prediction, , and control. The module employs models to process inputs and detect objects, including vehicles, pedestrians, cyclists, and elements, generating a 360-degree environmental understanding up to 400 meters ahead. The prediction module forecasts the trajectories and behaviors of dynamic agents, such as other vehicles and pedestrians, to anticipate potential interactions. The module optimizes safe and efficient paths by incorporating high-definition maps, traffic rules, and predicted scenarios. Finally, the control module translates these plans into precise vehicle commands for steering, acceleration, and braking via integration with the vehicle's APIs. The platform evolved through iterative advancements, with significant emphasis placed on simulation-based training by 2020 to accelerate development and testing. Argo AI's simulations enabled the logging of approximately 1 million virtual miles per day, accumulating billions of simulated miles overall to validate algorithms in diverse scenarios without real-world risks. This approach complemented on-road , allowing for rapid iteration on components across the stack. A innovation was the release of the Argoverse open-source dataset in 2019, comprising over 300,000 tracked driving scenarios collected from Argo AI's fleet in and . Designed to advance motion forecasting, the dataset includes 3D point clouds, camera images, high-definition maps with 290 km of annotated lanes, and trajectory annotations, facilitating research in behavior prediction, 3D object tracking, and scenario simulation for improved HD mapping. Argoverse has been widely adopted in the research community, supporting the development of more robust prediction models essential to the AV stack. To ensure scalability, the AV stack adopts a modular integration , allowing to various platforms while maintaining core functions. This design prioritizes hardware-agnostic software components, enabling deployment across different OEM architectures for urban Level 4 operations. The platform leverages infrastructure for efficient processing of vast data volumes from testing fleets, enabling model training, validation, and over-the-air software updates to deploy improvements fleet-wide.

Hardware and sensors

Argo AI's autonomous featured a custom roof-mounted sensor pod designed to provide comprehensive environmental . This pod integrated multiple sensor types for and 360-degree coverage, enabling the vehicles to detect objects in urban and settings. The hardware emphasized durability and precision, supporting operations in diverse conditions through high-resolution imaging and robust localization systems. The primary sensing components included two Velodyne VLP-32C lidars mounted on the roof, delivering a combined 64-beam configuration with a 200-meter range and overlapping 40-degree vertical for full 360-degree mapping. These spinning lidars generated approximately 107,000 points per sweep at 10 Hz, facilitating detailed 3D point clouds for obstacle detection and scene reconstruction. Complementing the lidars were units that provided and range data in all weather conditions, enhancing reliability for dynamic object tracking. The camera array consisted of seven high-resolution ring cameras (2048 x 1550 pixels at 20 Hz) arranged for panoramic 360-degree coverage and two front-facing cameras of the same resolution, positioned to capture depth information and fine visual details. These cameras featured improved in later iterations to better handle varying lighting and contrast, contributing to robust object . Inertial measurement units () and GPS receivers were integrated for 6-degree-of-freedom localization, achieving centimeter-level accuracy when fused with high-definition maps. Argo AI deployed this hardware on modified production vehicles, starting with third-generation Ford Fusion Hybrid sedans in 2019, which incorporated upgraded cameras and radars within the sensor pod. By 2020, the fourth-generation platform shifted to the Hybrid, refining the array for enhanced computing integration and sensor calibration. For commercial applications like delivery, the suite was adapted to Ford Transit vans, maintaining the roof-mounted configuration to support payload variations while preserving perception capabilities. Over time, Argo AI evolved its hardware by transitioning from third-party Velodyne lidars to in-house developed sensors, aiming for longer-range detection up to 400 meters for low-reflectivity targets to enable highway autonomy. This progression focused on cost efficiency and integration through supplier partnerships, while the overall design prioritized to ensure safe operation in adverse weather via multi-modal sensing. The sensor data was processed by onboard software for real-time fusion, though hardware robustness remained key to system performance.

Operations

Road testing programs

Argo AI initiated its road testing program in , , in August 2017, leveraging the city's varied urban environments to validate its autonomous driving platform on public roads. The company equipped modified vehicles with its self-driving system, focusing on data collection for perception, planning, and control algorithms while adhering to safety guidelines similar to those from the (DMV), including tracking disengagements where human operators intervened. Early testing emphasized geofenced operations within the city, with test specialists monitoring performance to ensure safe navigation through complex scenarios like intersections and pedestrian interactions. The program expanded in 2019 to include Miami, Florida, to evaluate performance in diverse weather conditions such as heavy rain and high humidity, which challenge sensor reliability and vehicle dynamics. That same year, testing extended to Detroit, Michigan, for cold-weather validation, including snow and ice impacts on traction and visibility, using the city's infrastructure to simulate winter urban driving. Additional U.S. sites included Palo Alto, California, starting in 2019 for Silicon Valley tech ecosystem integration and regulatory testing under California DMV oversight, and Washington, D.C., to assess performance in a dense federal district with unique traffic patterns. In September 2019, Argo AI began operations in Austin, Texas, targeting suburban and mixed-use scenarios with highways, residential areas, and construction zones to gather data on long-range routing and adaptability. These locations, along with others, allowed for broader operational design domain (ODD) coverage, with vehicles accumulating miles under supervised conditions to refine edge cases like sudden lane changes or emergency vehicle responses. Safety protocols formed the core of Argo AI's testing framework, incorporating two test specialists per vehicle—one as the primary operator and another for monitoring—along with remote human operators providing real-time guidance via over-the-air communication. Operations remained geofenced to predefined areas, with mandatory incident reporting to authorities and internal reviews of all disengagements, which were categorized as voluntary (operator preference) or mandatory (e.g., encounters with school buses). The company complied with and ISO 21448 standards for and , implementing fallback maneuvers to minimal risk conditions like safe stops. Public tests reported zero at-fault crashes attributable to the autonomous system; a 2018 Pittsburgh incident involved a running a red light and striking an Argo vehicle from behind, with no fault assigned to the self-driving technology. By May 2022, Argo AI achieved a key milestone with driving in select geofenced areas of and Austin, deploying fourth-generation vehicles without safety drivers aboard during daytime business hours to test complex urban maneuvers such as unprotected left turns and yielding to pedestrians. These driverless operations built on prior supervised testing, incorporating the company's hardware suite—including , , and cameras—for 360-degree perception, while remote monitoring ensured rapid intervention if needed. The extensive real-world data from these programs, including over 36,000 autonomous miles in alone during 2021 with minimal disengagements, informed iterative improvements to handle and enhance system robustness ahead of potential commercial deployment.

Commercial partnerships

Argo AI established key commercial partnerships with automakers and ride-hailing providers to integrate its autonomous driving platform into real-world services, focusing on operations and delivery applications. In July 2021, Argo AI, alongside Ford, announced a collaboration with to deploy self-driving hybrid vehicles on the ride-hailing network, beginning with pilot programs in later that year and expanding to Austin in 2022. These initial deployments featured safety drivers, with plans for fully driverless services targeted for 2023 across multiple cities, aiming to scale to at least 1,000 vehicles. As part of the agreement, acquired a 2.5% equity stake in Argo AI, reflecting the company's strategy of exchanging technology access for ownership interests in mobility operators. In the delivery sector, Ford and Argo AI partnered with Walmart in September 2021 to pilot autonomous last-mile delivery services using customized Ford Transit vans equipped with Argo's technology. The trials, which commenced in 2022, operated in designated areas of Miami, Austin, and Washington, D.C., transporting groceries and other goods to customers' doorsteps to demonstrate scalable urban logistics. This initiative highlighted Argo AI's role in enabling commercial applications beyond passenger transport, with Walmart leveraging the platform to enhance efficiency in e-commerce fulfillment. Argo AI's partnership with , formalized through a $2.6 billion in June 2020, extended to commercial deployments in , including the integration of its platform into the ID. Buzz electric van for autonomous operations. Testing of ID. Buzz prototypes began in in summer 2021, marking the first international trials of the jointly developed Level 4-capable vehicle, with additional road testing in to prepare for urban ride-pooling scenarios. Prior to Argo AI's shutdown in October 2022, the collaboration advanced plans for a service in , where modified ID. Buzz vehicles were slated for commercial launch by 2025. A key element of this alliance involved joint efforts with MOIA, the automaker's ride-hailing subsidiary, to pioneer autonomous ride-pooling in . Announced in 2021, the initiative outlined MOIA as the inaugural operator of an ID. Buzz-based fleet for on-demand services in , building on MOIA's existing electric ride-pooling network established since 2018. This venture aimed to create a seamless mobility ecosystem, with Argo AI providing the core self-driving software and hardware suite. Argo AI's revenue model centered on licensing its autonomous platform to strategic partners while securing equity stakes in downstream service providers to align incentives for commercialization. Through deals like the one with , the company generated income from technology deployment fees and shared upside in ride-hailing operations, avoiding direct fleet ownership to focus on scalable software provision. This approach, evident in partnerships with Ford and , positioned Argo AI as a B2B enabler of AV services rather than a consumer-facing operator.

Research collaborations

University partnerships

Argo AI established its primary university partnership with (CMU) in 2016, leveraging the expertise of its founders, who were alumni of CMU's Robotics Institute, to create a talent pipeline and shared research facilities in . This collaboration expanded in 2017 to include joint research initiatives with CMU faculty. In 2018, Argo AI committed $15 million over five years to establish a dedicated research lab at CMU, formally announced in 2019 as the Carnegie Mellon University Argo AI Center for Autonomous Vehicle Research, focused on collaborative R&D in perception algorithms and related technologies to enhance autonomous vehicle safety. Argo AI also formed agreements with the Georgia Institute of Technology in 2017 to support faculty-led research on self-driving technologies, contributing to projects like the Argoverse dataset for mapping and motion forecasting. Additional partnerships included the for advancing sensor perception and motion prediction, as well as Caltech and the for broader autonomous systems research. To support talent development, Argo AI implemented and hiring programs drawing from these partner universities, fostering a with strong academic ties in and AI. These alliances enabled broader technological contributions in autonomous driving through shared expertise and resources.

Academic contributions

Argo AI significantly advanced autonomous vehicle through the release of the Argoverse datasets, which provided large-scale, annotated for perception and tasks. In 2019, the company launched Argoverse 1.0, featuring over 300,000 scenarios—including 324,557 trajectories for motion and 113 scenes for 3D tracking—paired with high-definition maps to support global studies in motion and object tracking. This dataset, collected from diverse urban environments in and , facilitated the development of algorithms for understanding dynamic driving scenes. Building on this foundation, Argo AI introduced Argoverse 2.0 in 2022, expanding to 1,000 annotated scenarios with extensive 3D object tracking annotations and 250,000 motion forecasting examples from six U.S. cities, including Austin, , , Palo Alto, , and Washington, D.C. The dataset incorporated multimodal sensor data—lidar, stereo cameras, and ring cameras—along with enhanced annotations for , enabling broader research in and forecasting under varied conditions. These releases, supported by university partnerships, democratized access to high-quality autonomous driving data for the academic community. Argo AI researchers co-authored influential papers at major conferences, including the Argoverse works at CVPR 2019 and NeurIPS 2022 Datasets and Benchmarks track, which detailed the datasets' construction and benchmarks. Additional contributions included explorations of neural radiance fields () for high-fidelity mapping, such as depth-supervised variants that improved 3D scene reconstruction from sparse views, aiding virtual simulations in autonomous systems. The company open-sourced supporting tools, including the Argoverse on , which allows researchers to load, visualize, and simulate scenarios from the datasets for virtual testing; this repository has been utilized by thousands of developers worldwide. Argo AI also contributed to autonomous vehicle benchmarks by providing comparative annotations and evaluation metrics that complemented datasets like nuScenes, enhancing standards for in multi-object tracking. By 2023, the original Argoverse paper had amassed over 1,000 citations on , underscoring its impact on motion prediction research and inspiring subsequent works at startups like Aurora, which adopted similar open-data strategies for advancements. Following Argo AI's shutdown in 2022, the Argoverse team has continued to support and expand the datasets, including updates and hosting challenges at conferences such as the 2024 CVPR Workshop on Autonomous Driving.

Legacy and impact

Industry influence

Argo AI played a significant role in accelerating the of autonomous vehicle (AV) hardware, particularly through its development of in-house lidar sensors designed for cost-effective, large-scale . By 2021, the company unveiled a long-range capable of detecting objects up to 400 meters away, which addressed key limitations in resolution and range while enabling broader adoption by original equipment manufacturers (OEMs). This effort contributed to industry-wide pressure on suppliers, including Velodyne, to lower prices; for instance, Velodyne reduced the cost of its popular VLP-16 sensor by 50% in 2018, aligning with broader trends where automotive lidar unit costs dropped approximately 90% since 2015 due to increased demand and technological advancements from players like Argo AI. The company also advocated for robust regulatory frameworks to ensure AV safety, participating in industry efforts to shape federal policies during its operational years. In 2022, Argo AI established the Argo Safety Advisory Council, comprising external experts who advised on safety practices and building in self-driving technology, influencing discussions on national standards for automated systems. This advocacy aligned with broader U.S. ional examinations of AV safety, such as hearings in 2019 and 2021 that emphasized the need for coordinated federal guidance on testing and . Argo AI's operations fostered talent migration within the AV sector, with former employees contributing expertise to various companies and initiatives, thereby elevating overall industry capabilities. The company's Pittsburgh headquarters helped solidify the city as an AV hub, attracting over $3 billion in tech investments by 2019, including Ford's $1 billion commitment in 2017 and a combined $3.6 billion from Ford and , which spurred local job creation and ecosystem growth with annual compensation exceeding $1.2 billion. Furthermore, Argo AI demonstrated the viability of models for OEM-shared AV development, as evidenced by its structure backed by Ford and , which secured over $7 billion in valuation by and enabled collaborative scaling of Level 4 technology for urban ride-hailing and delivery. This approach proved effective in pooling resources and expertise before OEMs shifted toward in-house efforts, setting a precedent for efficient, multi-stakeholder innovation in the sector.

Post-shutdown developments

Following the shutdown of Argo AI in late 2022, Ford integrated key technologies and personnel from the company into its new subsidiary, Latitude AI, launched in March 2023. Latitude AI hired approximately 550 former Argo AI employees specializing in areas such as , , mapping, and sensors to advance Ford's advanced driver-assistance systems (ADAS). This included applying Argo's mapping expertise to enhance the BlueCruise hands-free driving system, which by 2023 supported highway autonomy across more than 130,000 miles of pre-mapped roadways in , enabling features like eyes-off driving on compatible highways. As of 2025, Latitude AI continues to develop Level 3 automated driving systems for Ford, with deployment anticipated in 2026. Volkswagen, through its software subsidiary Cariad, absorbed select Argo AI projects and provided continued employment opportunities for many of the startup's employees to support ongoing autonomous driving development. Cariad focused on integrating promising Argo technologies into its internal efforts for highly automated driving systems, aligning with 's shift toward software-defined vehicles and partnerships like those with Bosch for Level 2+ and Level 4 capabilities. By 2023, Cariad advanced features such as enhanced Travel Assist in models like the ID.7, which includes assisted lateral and longitudinal control for highway driving. Argo AI alumni have contributed to new ventures in the autonomous vehicle space post-shutdown. Notably, co-founders Bryan Salesky and Pete Rander launched Stack AV in September 2023, a Pittsburgh-based startup focused on AI-powered autonomous trucking systems to optimize supply chains and address safety challenges in freight transportation. Backed by SoftBank, Stack AV employed around 150 people as of 2023, growing to approximately 200 by late 2025, many with Argo AI experience. Intellectual property from Argo AI was transferred following the closure, with 137 U.S. patents related to and autonomous driving systems assigned to Ford Global Technologies LLC in March 2023. An additional 83 patents on and imaging were transferred to Co. Ltd. in April 2023. These transfers supported the redistribution of Argo's assets to sustain advancements in and technologies at Ford and other partners.

References

  1. https://www.theverge.com/2017/8/16/16155254/argo-ai-ford-self-driving-car-autonomous
  2. https://www.crunchbase.com/organization/argo-ai
  3. https://pitchbook.com/profiles/company/170482-60
  4. https://www.theverge.com/2022/10/26/23423998/argo-ai-shut-down-ford-vw-av-self-driving
  5. https://www.sunsethq.com/blog/why-did-argo-ai-fail
  6. https://www.reuters.com/technology/ford-volkswagen-backed-argo-ai-shut-down-techcrunch-2022-10-26/
  7. https://techcrunch.com/2022/10/26/ford-vw-backed-argo-ai-is-shutting-down/
  8. https://technical.ly/startups/stack-av-argo-ai-founders-launch/
  9. https://www.crunchbase.com/organization/stack-av
  10. https://media.ford.com/content/fordmedia/feu/de/en/news/2017/02/10/ford-invests-in-argo-ai-new-artificial-intelligence-company.html
  11. https://www.geekwire.com/2017/ford-argo-google-uber-ai-autonomous/
  12. https://www.wired.com/2017/02/eyeing-self-driving-future-ford-drops-1b-ai-startup/
  13. https://www.bloomberg.com/news/features/2018-09-25/this-driverless-startup-has-one-mission-save-the-future-of-ford
  14. https://www.theinformation.com/articles/argo-ai-takes-different-route-to-self-driving-cars
  15. https://www.forbes.com/sites/samabuelsamid/2020/06/02/ford-volkswagen-argo-investment-deal-closes-argo-ai-expands-to-europe/
  16. https://www.reuters.com/business/autos-transportation/ford-vw-pop-automated-vehicle-bubble-with-argo-ai-exit-2022-10-26/
  17. https://media.ford.com/content/fordmedia/fna/us/en/news/2022/10/26/ford-to-advance-on-autonomous-and-connected-vehicle-tech.html
  18. https://www.cnn.com/2022/10/26/business/ford-argo-ai-vw-shut-down
  19. https://www.volkswagen-group.com/en/press-releases/volkswagen-focuses-development-for-autonomous-driving-16777
  20. https://arstechnica.com/cars/2022/10/argo-ai-will-cease-operations-as-ford-and-volkswagen-pull-investments/
  21. https://annualreport2022.volkswagenag.com/group-management-report/results-of-operations-financial-position-and-net-assets/results-of-operations.html
  22. https://www.autoweek.com/news/technology/a41789388/argo-ai-autonomous-developer-shutdown/
  23. https://www.theverge.com/22627847/argo-ai-bryan-salesky-decoder-interview-lyft-self-driving
  24. https://www.eetasia.com/cooperation-competition-dynamics-across-av-software-platforms/
  25. https://arxiv.org/abs/1911.02620
  26. https://www.prnewswire.com/news-releases/argo-ai-introduces-complete-line-of-autonomous-vehicle-products-and-services-to-serve-rideshare-and-delivery-301622262.html
  27. https://www.argoverse.org/av2.html
  28. https://corporate.ford.com/articles/products/fords-next-gen-test-vehicle.html
  29. https://www.forbes.com/sites/samabuelsamid/2021/05/04/argo-ai-shares-some-details-on-the-lidar-that-will-be-used-on-ford-automated-vehicles/
  30. https://aimagazine.com/technology/argo-ai-creating-safe-autonomous-vehicle-technology
  31. https://www.carscoops.com/2019/06/ford-and-argo-ai-launch-their-third-gen-autonomous-fusion-hybrid/
  32. https://media.ford.com/content/fordmedia/feu/en/news/2021/06/17/-would-you-trust-your-parcels-to-a-self-driving-van--ford-and-he.html
  33. https://techcrunch.com/2021/05/04/argos-new-lidar-sensor-could-help-ford-vw-deploy-self-driving-vehicles-at-scale/
  34. https://www.northlandinjurylaw.com/blog/argo-self-driving-vehicle-involved-recent-pittsburgh-crash/
  35. https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-vehicles/disengagement-reports/
  36. https://fleetwiki.net/documents/ArgoSafetyReport.pdf
  37. https://editions.mydigitalpublication.com/article/TRANSPORTATION/3446226/607478/article.html
  38. https://techcrunch.com/2019/06/12/argo-ai-is-taking-its-new-generation-of-self-driving-vehicles-into-downtown-detroit/
  39. https://techcrunch.com/2021/07/30/argo-ai-can-now-offer-the-public-rides-in-its-autonomous-vehicles-in-california/
  40. https://www.reuters.com/business/autos-transportation/argo-launches-non-commercial-driverless-vehicles-miami-austin-2022-05-17/
  41. https://www.cnbc.com/2019/09/25/ford-selects-austin-to-test-deploy-autonomous-vehicles.html
  42. https://techcrunch.com/2022/05/17/argo-ai-driverless-autonomous-vehicles-miami-austin/
  43. https://www.forbes.com/sites/samabuelsamid/2022/05/17/argo-ai-begins-driverless-testing-in-austin-and-miami/
  44. https://www.reddit.com/r/SelfDrivingCars/comments/somrjp/disengagement_reports_for_2021/
  45. https://techcrunch.com/2021/07/21/argo-ford-to-launch-self-driving-vehicles-on-lyfts-ride-hailing-app/
  46. https://www.reuters.com/business/autos-transportation/argo-ai-ford-lyft-launch-self-driving-ride-hail-service-miami-austin-2021-07-21/
  47. https://www.cnbc.com/2021/09/15/walmart-to-test-self-driving-cars-in-multiple-cities-with-ford-argo.html
  48. https://www.supplychaindive.com/news/walmart-ford-argo-ai-autonomous-vehicles-delivery/606659/
  49. https://www.freightwaves.com/news/walmart-drives-closer-to-autonomous-delivery-with-ford-argo-ai-deal
  50. https://www.zdnet.com/article/argo-ai-closes-2-6bn-volkswagen-investment-in-european-us-push-with-help-from-ford/
  51. https://media.vw.com/releases/1532
  52. https://www.forbes.com/sites/alanohnsman/2021/09/05/volkswagen-argo-ai-ready-electric-robotaxi-vans-for-german-test-fleet/
  53. https://www.moia.io/en/news-old/vwcv-moia-and-argo-ai-present-roadmap-for-autonomous-ride-pooling-in-hamburg
  54. https://www.itsinternational.com/its4/its7/news/av-ride-pooling-hamburg-2025
  55. https://fleet.ie/volkswagen-commercial-vehicles-argo-ai-moia-show-first-id-buzz-prototype-for-autonomous-driving/
  56. https://www.forbes.com/sites/samabuelsamid/2021/07/21/lyft-invests-in-argo-ai-ford-robotaxis-in-late-2021/
  57. https://www.nydailynews.com/2021/07/21/ford-argo-ai-team-with-lyft-to-launch-self-driving-robotaxi-service-in-miami-austin/
  58. https://www.axios.com/2017/12/15/ford-invests-1-billion-in-autonomous-driving-startup-argo-ai-1513300408
  59. https://www.forbes.com/sites/doronlevin/2019/06/25/argo-ai-self-driving-deal-with-carnegie-mellon-aims-to-erase-uber-disappointment-of-2016/
  60. https://www.ri.cmu.edu/argo-ai-collaborates-with-cmu-georgia-tech-researchers-on-self-driving-cars/
  61. https://www.post-gazette.com/business/tech-news/2018/09/06/argo-ai-self-driving-autonomous-vehicles-pittsburgh-peter-rander-bryan-salesky/stories/201809060044
  62. https://techcrunch.com/2019/06/24/argo-ai-is-investing-15-million-into-a-self-driving-car-research-center-at-cmu/
  63. https://www.cs.cmu.edu/news/2019/carnegie-mellon-university-and-argo-ai-form-center-autonomous-vehicle-research
  64. https://argoverse.org/about.html
  65. https://www.autonomousvehicleinternational.com/features/interview-brett-browning-cto-argo-ai.html
  66. https://argoverse.org/av1.html
  67. https://openaccess.thecvf.com/content_CVPR_2019/html/Chang_Argoverse_3D_Tracking_and_Forecasting_With_Rich_Maps_CVPR_2019_paper.html
  68. https://argoverse.org/av2.html
  69. https://arxiv.org/abs/2301.00493
  70. https://github.com/argoverse/argoverse-api
  71. https://argoverse.org/sceneflow.html
  72. https://argoverse.org/E2EForecasting.html
  73. https://www.eetimes.com/velodyne-cuts-price-popular-lidar-sensor-50/
  74. https://innoviz.tech/blog/understanding-the-factors-that-influence-the-price-of-lidar
  75. https://www.theverge.com/2022/8/4/23292533/argo-ai-safety-experts-advise-autonomous-vehicle
  76. https://www.repairerdrivennews.com/2022/08/24/steps-taken-toward-av-safety-testing-regulation-by-argo-ai-state-federal-legislators/
  77. https://www.congress.gov/event/116th-congress/house-event/110513
  78. https://www.govinfo.gov/content/pkg/CHRG-117hhrg47612/html/CHRG-117hhrg47612.htm
  79. https://nextpittsburgh.com/latest-news/2019-was-a-huge-year-for-pittsburgh-startups-with-nearly-3-billion-in-investment-funding/
  80. https://www.wesa.fm/economy-business/2022-10-26/a-once-hot-startup-pittsburgh-based-argo-ai-will-soon-be-no-more
  81. https://ridc.org/wp-content/uploads/2021/09/PGH-Autonomous-Systems-Full-Report-1.pdf
  82. https://media.ford.com/content/fordmedia/feu/en/news/2019/07/12/ford-_-volkswagen-expand-their-global-collaboration-to-advance-a.html
  83. https://www.forbes.com/sites/jimcollins/2019/07/12/vws-investment-in-fords-argo-ai-comes-with-cash-but-without-a-projected-launch-date-for-avs/
  84. https://media.ford.com/content/fordmedia/feu/en/news/2023/03/02/ford-establishes-latitude-ai-to-develop-future-automated-driving.html
  85. https://www.ford.com/electric/ev-news/ev-expertise/ford-bluecruise?intcmp=evNews-cta-expertiseBluecruise
  86. https://fordauthority.com/2025/10/ford-ceo-jim-farley-says-blue-oval-wont-be-first-to-roll-out-l3-highway-driving/
  87. https://www.volkswagen-newsroom.com/en/the-volkswagen-id7-17897/operating-principle-of-the-id7-assist-systems-17902
  88. https://www.therobotreport.com/argo-ai-founders-launch-autonomous-trucking-startup-stack-av/
  89. https://www.reuters.com/business/autos-transportation/softbank-backs-new-autonomous-trucking-startup-founders-defunct-argo-ai-2023-09-07/
  90. https://bitscale.ai/directory/stack-av
  91. https://www.linkedin.com/company/stackav
  92. https://www.inquartik.com/blog/top-patent-transactions-2023-blackberry/
Add your contribution
Related Hubs
User Avatar
No comments yet.