“Ride with Pony.ai”

Founded: 2016 in Fremont, USA

Category: Automotive, Autonomous Vehicles, Transportation, Artificial Intelligence

Primary office:Fremont, USA; Guangzhou, China

Core technical team: San Francisco bay area, USA

Status: Private

Employees:

251-500

Amount raised: USD $726 million (5 rounds)

OVERVIEW

Develops AI applications in the field of autonomous robotics with a main focus on reliable self-driving technologies. In particular, the company specializes in Level 4 autonomous driving technology for passenger cars and trucks.

PERFORMANCE METRICS

  • Market cap or value of company: USD $3 billion (February 2020)
  • Revenue: USD $10-50 million (2019)

ACHIEVEMENTS

  • A strategic cooperation with the Shanghai government on vehicle development in the Yangtze River Delta Urban Agglomerations and a Robo Taxi operation in Shanghai (July 2020)
  • ai reaches a post-money valuation in the range of $1B to $10B (February 2020)
  • Reaches 1 Million Kilometers Milestone (November 2019)
  • Partnering with Japanese auto market leader Toyota in Beijing and Shanghai in the area of ‘safe’ mobility services (August 2019)
  • Pony.ai begins working on truck technologies (April 2019)

Sells

  • RoboTaxi –  a self-driving taxi service
  • Level 4 (pilot) cars – vehicles able to operate without human oversight under select conditions
  • Self-driving platforms – software and technology system components

Channels

  • Several strategic agreements/partnerships with global automotive giants: 1) Toyota [Pony.ai’s pilot program in Beijing and Shanghai (‘safe’ mobility services)]; 2) Hyundai [Establishing a self-driving system for pilot ride-sharing service BotRide and competing against Google’s Waymo].
  • Developing strategic partnerships with critical technology suppliers: 1) GAC Group (a Guangzhou-based automobile maker), to develop level 4 robo-taxi vehicles; 2) On Semiconductor, to prototype image sensing and processing technologies for machine vision.
  • Building relationships with local governments and policy makers. For example: 1) State of California, for obtaining a road test license in June 2017; 2) Beijing Municipal Government, for obtaining the highest level of road test license T3 in 2018

Competencies

  • Developing “geofenced technologies” – mapping driving areas into a full map of city zones that are safe for autonomous cars.
  • Road tests to simulate driving conditions
  • The autonomous driving platform powers the first fully self-driving fleet in China.
  • The full-stack hardware platform, PonyAlpha, leverages lidars, radars, and cameras to keep tabs on obstacles within 200 meters of its self-driving cars. The platform serves as the foundation for the company’s fully autonomous trucks and freight delivery solution.

Distinct AI Features

Type

  • Not specifically mentioned. The company uses mainly simulation models to test AI algorithms.
  • Applications include 1) computer vision (gauging driver attention for handoff challenges with assisted driving systems), 2) machine learning (enabling software-powered devices to collect, process and analyze data locally, without resorting to the cloud), and 3) AI and deep learning technologies.

  AI use

  • The goal of the company is to bring together all disparate components to a fully-autonomous, level-four vehicle platform. Such components include hardware and software modules such as sensors, sensor fusion processors, and simulation models to test AI algorithms.
  • To constrain engineering challenges, Pony.ai has focused on building an autonomous car that is restricted to more predictable environments using AI technology. Such environments include college campuses, suburban towns, industrial settings, and other environments where the complexity of driving is significantly less.
  • The focus of the company is on using AI as part of the vehicle system. AI and deep learning technologies are used to assess road conditions, other road users and driver behavior to power in-vehicle safety alerts.

AI useRate of return on customer’s investment to make AI work

Immediate:

  • Diversifying the range of services available to meet business objectives

Long term:

  • Higher profitability by decreasing transportation costs and reducing delivery inefficiencies

Databases

  • Data collected from connected vehicles to inform better the AV design, including how drivers respond to various situations on the road

Quantum Computing

  • None identified.

Resources

Assets

  • Talent investment in a strong software engineering team
  • ai is led by ex-Baidu chief architect James Peng and Tiancheng Lou (an ex-Google engineer who worked at Google X’s autonomous car project)
  • Patent technologies
  • Strong financial support

Processes

  • Government outreach: Building government relationships in regions of interest for future operations
  • Simulations: Pony.ai works on training its systems on roads in a closed site; in-vehicle safety procedures; and some other 39 different features related to having a working self-driving system without critical flaws.
  • Developing technologies to quickly synchronize and interpret information gathered by different sensors on self-driving cars

Priorities

  • Constant pursuit of technology to increase the reliability of Pony.ai’s software and hardware systems
  • Improving the efficiency of the fleet operation and scalability of technology
  • Taking the lead in managing uncertainty in open driving environment
  • Continue expanding testing and training activities in multiple locations around the globe
  • Maintaining an outpost in California for attracting engineering talent, easier access to financing, and access to a good road infrastructure

References

  1. Crunchbase. 2020. Pony.AI. www.crunchbase.com/organization/pony-ai.
  2. Pony.AI website. (2020). https://www.pony.ai/en/index.html.
  3. Wiggers, K. (2020). Driverless car startup Pony.ai raises $462 million at a $3 billion valuation https://venturebeat.com/2020/02/25/driverless-car-startup-pony-ai-raises-462-million-at-a-3-billion-valuation/.

Contributors

  • Radomir Todorov