Autonomous Vehicle Ecosystem Analysis & Opportunities

Autonomous Vehicle Ecosystem Analysis & Opportunities April 2019

Source : DRAUP 1 AGENDA

This section provides an overview of : 01 Autonomous Vehicle Overview  Autonomous vehicle overview and its potential 02 Technology Spend Analysis  Disruption in Autonomous Vehicles: o Tech Giants 03 Autonomous Vehicle Adoption o Partnerships & Consortiums o Acquisitions & Start-ups  Impact of AV on different industries 04 Bay Area–Deep Dive  Future of AV  Growth Drivers of AV  New Emerging business models 05 Top Companies Deep Dive

06 Partnership Opportunities

2 Overview: Autonomous Vehicles (AV) have huge potential to impact global economies, markets and industries $7 Trillion Potential savings in the areas of fuel efficiency, cost of life and productivity gains enabled through AV based business models in US by 2025 $250 Billion Estimated worth of Autonomous Vehicle Industry by 2025

3 Million Potential Job loss in US by 2025 17% Expected AV market share as percentage of total worth of Auto industry, in 2025

8 Million Estimated Level 3 and higher AV by 2025 3

Note : DRAUP- The platform tracks engineering insights in the automotive ecosystem using our proprietary machine learning algorithms along with human curation. The platform is updated in real time and analysis is updated on a quarterly basis Source : DRAUP 3 Disruption in AV: Penetration of Tech giants in AV space has created an intense competition for traditional automotive players that enables multiple disruptions in the ecosystem

Disruption in AV Case Studies

Tech Giants Tech giants are penetrating the AV ecosystem due to its 1 Penetration prominent potential and impact. Tech giants are in AV investing with OEMs, Tier-1 suppliers and start-ups to offer services and solutions • Google’s Waymo, self driving vehicle technology company has partnered with Tier- 1 and OEMs like Magna, FCA, and Jaguar to offer full-stack autonomous vehicles.

• It is also setting up a factory in Detroit to build autonomous vehicles and is working with American Axle & Manufacturing to convert the existing factory Autonomous vehicle development has disrupted the traditional partnership trends in auto industry and has 2 Consortium & Partnerships brought in multiple industry giants together working in • Intel acquired Mobileye which develops computer vision and machine learning, consortiums data localization, localization and mapping for ADAS and autonomous driving.

• BMW collaborated with Intel and Mobileye to position itself in AV ecosystem. Followed by BMW, tier 1 suppliers and other OEMs like Delphi, Valeo, Magna, Toyota, Aptiv, Continental, Jaguar, and Audi have also joined the coalition.

AV based ML and Sensor startups have attracted 3 Acquisitions of phenomenal investments from the giants who are Start-ups looking to win the Autonomous vehicle race • GM acquired Cruise to use the technology and talent to accelerate the process of developing AV. GM Cruise is also partnering with other startups and companies to deploy autonomous vehicles. It has collaborated with DoorDash which offers food delivery service 4

Note : The platform tracks real time insights and developments in the Autonomous Vehicle Ecosystem such as global engineering footprint, product launch, Leadership Announcements, M&A, among other essential insights Source : DRAUP 4 Above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April, 2019 Penetration of industries in AV: New age solution providers in the areas of Semicon giants, Telecom, Cloud, and Mobility are bolstering the evolution of vehicle autonomy

• Traditional suppliers such as Automotive Ecosystem has been disrupted through digital mega innovations Ecosystem maturity trend during last 15 years Bosch and TomTom have enabled advanced vehicle navigation and Telecom: 5G Infrastructure 10 monitoring through specialised telematics equipment Insurance Providers: Usage based Insurance 9 Smart • Semiconductor giants such as Intel Mobility and Nvidia have developed Cloud Platforms: Data Management & Security 8 specialised SoCs for processing and computing large amount of New Age Suppliers: ADAS Systems & Components 7 vehicle datasets • Tech Mafia have transformed the Data Services: Connected Car 6 vehicle into a software computing system with capabilities to take Internet autonomous decisions Age Mobility Services: Alternative Ownership 5 • New age suppliers have built Tech Mafia: Car OS, HMI 4 capability into Advanced vehicle control using deep learning, sensor systems and connectivity 3 Consumer Electronics: Infotainment OS services Silicon Ecosystem Automotive in the players of Number Evolution 2 • The current Autonomous Vehicle Semiconductor Giants: SoC Processors ecosystem has been rapidly growing through a rich infrastructure of 1 network, cloud & insurance Automotive 1.0 Traditional Suppliers: Telematics equipment Note: Each unit on Y-Axis represents a providers enabling new age single type of ecosystem player 0 business models 2003 2006 2009 2012 2015 2019 5 Note: The timeline above is illustrative of landmark events in the autonomous vehicle ecosystem during the last 15 years. The list above is non exhaustive DRAUP Engineering Module: The platform tracks real time insights and developments in the Autonomous Vehicle Ecosystem such as global engineering footprint, product Source : DRAUP 5 launch, Leadership Announcements, M&A, among other essential insights Future of AV: Companies are accelerating commercialization of level 3 & 4 autonomy to lead the technology race

Targeted Levels of automation by 20212

Fully automated 5 Uber BMW General Motors Ford Waymo vehicle Full  The league of 5 are well Automation positioned and future-ready, basis their current R&D investment or Delphi via virtue of their acquisitions and/or partnerships Highway autopilot Apple Autoliv Volvo Argo.ai Valeo Daimler Bosch Tesla Intel Including highway 4 High  GM, Ford, and Waymo have Convoy Parking Automation committed to attain Level 5 PSA nuTonomy garage pilot Volkswagen automation capabilities whereas Intel, Tesla and Bosch have envisioned Level 4 automation by Highway chauffeur Nissan-Renault Nvidia Toyota Baidu 2021 Traffic jam chauffeur 3 Conditional Continental Automation Zoox Automation Nauto  These players have been exploring a diverse set of GTM strategies such as partnerships Partial automated with mobility providers, fleet Parking Traffic jam 2 management and personal Partial ownership model to launch their assistance Automation first commercial Autonomous Vehicles by 2021

2021 AV Readiness Index1 6

Note : 1-2021 AV Readiness Index: Function of % R&D Talent in Autonomous Vehicle technology, External Acquisitions and Investment, patents and partnerships; 2- Function of current leadership Outlook and commitments for autonomous vehicle launch in 2021. Automation Levels as outlined by SAE updated as of 2019; Source : DRAUP 6 The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 AV Growth Drivers: Liberal government policies, technology advancement and ecosystem openness to co-innovate are the key enablers driving autonomous vehicle innovations

1 Technology advancement 2 Political, legal and social drivers Drivers for Decline in cost of computing and advancement State legislations related to autonomous vehicles in processing power have enabled processing Autonomous have gradually liberalised . In 2019, 29 states large volume and variety of data such as image, have introduced legislation related to autonomous voice, text, etc. vehicles in USA, allowing testing of autonomous Vehicle fleets under certain specified conditions

Advances in machine learning have allowed computer vision to compute unstructured data Extensive government investment in key and distinguish objects on the road and build 3-D countries- US and UK governments plan to invest maps of the surrounding area $4Bn and £38Mn over the next 5 years, on driverless cars technology

Deep learning and artificial intelligence have 3 Open Ecosystem led to better algorithms for pedestrian detection, Projected 20% overall reduction in road traffic control and other automation features. accidents- Elimination of drivers is expected to • Collaborative and open innovation- Top player Tesla reduce driving accidents caused by human error. open-sourced its patents while Baidu and Lyft have open software platforms

• Competitive landscape- Entrance of technology mafias which are building a competitive environment in AV through their strong capability in software platforms

• R&D partnerships between universities and automakers- Toyota partnered with University of Michigan for autonomous innovation. 7

Note: Autonomous Vehicle regulations have been verified from reports published by Department of Motor Vehicle, California and other state regulatory bodies in respective geographies The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on Feb, 2018 Source : DRAUP 7 New Business Models: Shared service model and fleet owned taxis would be the first level of AV integration globally

Service and public utilization based models to dominate while traditional ownership model to diminish

Business Model Description Examples Intensity of Autonomy

Privately owned vehicles provide Individual Owned Shared Service ride hailing/sharing service when Uber, Lyft Models Emerging owner is not currently using it. Models Service company operates fleet of Fleet Owned Taxis autonomous vehicles to provide Waymo, NuTonomy, Lyft mobility services

Consumers pay owner for the use Vehicle Licensing Customizable rental programs of vehicle Services and software that unlock Potential AV enabled software packages Productive software suites – full autonomous capabilities Models Package of Hardware and Software Retrofit to retrofit fully autonomous Comma One capabilities on selected vehicles

Note: Autonomous Vehicle models have been verified from reports published by Department of Motor Vehicle, California and other state regulatory bodies in respective geographies. The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on Feb, 2018 Source : DRAUP 8 AGENDA

This section provides an overview of : 01 Autonomous Vehicle Overview  Technology Spend Analysis: o In-house Engineering Spend 02 Technology Spend Analysis o External Technology Spend  Engineering Spend Analysis: o Top Companies 03 Autonomous Vehicle Adoption o Industry o Geography o Technology Segments 04 Bay Area–Deep Dive  Geographical Talent Analysis  AV Ecosystem Analysis and Top 25 Companies mapping 05 Top Companies Deep Dive  Analysis of acquisitions and investments by top companies

06 Partnership Opportunities

9 AV Technology Spend: AV ecosystem players are fuelling the technology spend by making investment in-house or externally for faster development and deployment of AV capabilities

Autonomous Vehicle In-house1 engineering spend • Engineering spend by autonomous vehicles includes in- house investments like talent, solutions, platforms, and services made by the OEMs, Tier-1suppleirs, tech giants to enhance the autonomous technologies for faster $10-11 Bn deployment of vehicles Engineering spend globally on • Major players investing in engineering spend include: autonomous technologies as of 2018 Total AV Technology Spend by top 25 players (2018): Autonomous Vehicle External2 $32 34 Bn technology spend – • Semicon giants, OEMs, tech giants are investing or acquiring in start-ups to leverage AV innovations. For example, Google acquired Waze, GM acquired Cruise, Intel acquired Mobileye to develop AV solutions and capabilities $22-23 Bn AV –Acquisition • Major players include acquiring or building consortiums: ,Corporate VC Spend & Partnership as of 2018

Note: AV In-house Technology spend: includes salaries and compensation along with spend on software, platforms and hardware tools required to develop In-house capability; External Technology Spend: Consists of investment in Autonomous Vehicle and related technology areas through Acquisitions, Partnerships and Corporate Venture Arms; Source : DRAUP 10 The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Engg Spend by Top Companies: Engineering spend by the top 5 players is largely focused on developing full stack solutions, robust sensor systems and advanced computing platforms for vehicle control

AV In-house Engineering spending analysis (2018) Key Insights

• Majority of the in-house engineering spend by players is ~56% being invested in Autonomy. Top companies have already R&D spend invested billions in development of autonomous vehicles by top 5 like Ford is investing $5.4 billion in driverless cars and GM players has already invested $1.5 billion. $10 11 • GM, Tesla, Ford, Waymo, and Uber are the top players in – ~31% the autonomous vehicles ecosystem and are highly Bn R&D spend investing to deploy the AV. Companies like GM and In-house by next 10 Waymo are building R&D centres and assembly plants to Engineering players build AV Spend by top 25 companies • Primarily, the companies are focusing on ride-sharing and delivery over individual ownerships.

• Top players are majorly focusing on Electric vehicle AV as ~13% compared to gasoline EV due to less moving parts and R&D spend maintenance costs by next 10 players

11 Note: The numbers above are rounded-off, so they might not add up to 100% Note: 1-Technology spend includes employee compensation and related expenses along with spend on software, platforms and hardware tools required to develop In-house capability; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 11 AV In-house Engg Spend by Industry: Tech Mafia and the Semiconductor giants are spending heavily alongside Automakers to develop strong Autonomous Vehicle capability

AV In-house Engineering Spend by Industry Verticals (2018) • OEMs have strategic focus on developing critical safety and driving systems in-house. OEMs such as Daimler, BMW and Ford are establishing partnerships with technology providers to collaboratively develop software capability for vision Traditional OEMs 34-36% and perception systems

• Semiconductor giants such as Intel and Nvidia have In-house developed specialised Autonomous Vehicle SoCs Engineering Tech Mafia 25-27% for processing and computing using ML algorithms spend1 on AV as of 2018 • Tech Mafia giants are differentiating through strong Semiconductor 14-16% AI capability leveraging deep learning algorithms $10–11 required to make advanced driving systems safe Bn and predictable Tier-1 Suppliers 10-12% • Tier-1 suppliers such as Bosch, Delphi and Continental are major players providing Sensor Systems such as Lidar, Radar, Cameras, and Ultrasonic sensors Automotive Start-ups 7-8% • Full stack ADAS providers is the most funded segment . Nauto, Argo AI and Drive.ai are the top Others* <5% players investing in full stack-Autonomous Vehicle solutions

Note: The numbers above are rounded-off, so they might not add up to 100% Others* include Telecom, Data Services, Insurance and other AV related infrastructure providers 12 Note: 1 Include investments on In-house R&D spend on engineering salaries and infrastructure support in AV and related technology areas; DRAUP Engineering Module – Include AV companies across major geographies such as US, Canada, Israel, Europe, China and India. Source : DRAUP 12 Coverage may be limited in China and other south east APAC regions AV In-house Engg spend by geography: Majority of engineering investment in AV ecosystem is consolidated in US due to the supporting regulations by NHTSA

USD 10-11 Bn Global AV In-house Engineering Spend by Top 25 players (2018) Investment focus by Geography

. OEMs, Tech giants, Tier-1 providers have chosen US, UK, Germany, China and India as major hotspots for engineering centres

. Majority of the players are focusing in US for the development of AV ~30% solutions due to support from the NHTSA and technological ~49% Europe advancement. US have also ~21% introduced regulations for self-driving vehicles on public roads and issued Americas autonomous testing permits. APAC California has allowed operation of fully autonomous vehicles with no driver on public roads

. Autonomous Vehicles spend in APAC region is growing due to high autonomy activities by Chinese players like Baidu and SAIC. Shanghai has issued its first self- driving licenses in China

Note: Geographical split indicates only the prime Autonomous Vehicle R&D locations. Primary locations include US, Europe, India and China; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 13 Geographical Talent Split: While the AV talent footprint is distributed across global locations, US and Europe are the hotspots with nearly 60% of talent consolidated between these two regions

Geographical split by AV Engineering Headcount

USA ~47%

Europe ~15%

Global Autonomous Vehicle China ~13% Engineering Headcount

40,000–45,000 UK ~9%

Israel ~7%

Canada ~6%

Singapore ~3%

14 Note: The numbers above are estimated R&D headcounts in respective locations updated as of 4th quarter of financial year 2018

Note: Geographical split indicates only the prime Autonomous Vehicle R&D locations. Primary locations include US, Germany, France, Canada, China, UK, and Singapore ; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 14 In-house Engg Spend by Technology Segments: R&D is focussed on developing core software capabilities, leveraging deep learning for computing, vehicle control and vision-based perception

Technology Segments INSIGHTS Total AV Engineering Spend1 USD $ 10-11 Bn 3D Mapping/ Localization High resolution HD Maps enable precise lateral and 11 % longitudinal positioning for vehicle localization

Vision based perception Computer Vision systems use advanced deep learning to aggregate, classify and identify critical environment data such as obstacles, pedestrians, traffic 24 % signs etc.

HMI is crucial to optimally support the driver in the HMI/ UI-UX monitoring and remotely control autonomous cars and to give access to live sensor data and useful data about 10 % the car state, such as current speed, engine and gear state

Computing & Vehicle Control Using Neural Networks, the vehicle brain analyses all sensor input and operates steering, accelerator and brakes for critical driving decisions such as collision 30 % warning, cruise control and advanced safety

Sensors Lidar, Radar, Odometry and Ultrasonic sensor systems for lane centering, path planning and V2V 20 % communications Network, Connectivity & Security Advanced vehicle connectivity infrastructure to enable communication between vehicles and environment (V2V, V2X) 5 % 15

Note: 1-Technology spend includes employee compensation and related expenses along with spend on software, platforms and hardware tools required to develop In-house capability; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 15 AV Ecosystem: Two type of organisations are accelerating Autonomous Vehicle Ecosystem - In-house Innovators vs Collaborative Developers

1 In-house Innovators Autonomous Vehicle Capability & Investment Analysis 1 In-house Innovators Collaborative Developers 2 • Start-ups and Tech Mafias have High been investing in Autonomous Zoox Automation Vehicle platforms and Vehicle Tesla Google- Waymo BMW Operating Systems, leveraging their > strong software capability nuTonomy Argo.AI GM-Cruise ------Daimler Intel • Key technology focus areas of

2 Baidu these companies are Deep Nauto Uber learning for vehicle control and Computer Vision for environment Volkswagen Valeo Volvo Toyota Ford perception and sensing PSA Delphi Apple Autoliv Bosch Nissan-Renault 2 Collaborative Developers • Semiconductor giants such as Nvidia Intel and OEMs such as BMW, Continental Toyota and GM have established Technology Maturity Index Technology strong consortium to co-innovate

• These players have also acquired many companies which offer full stack Autonomous Vehicle Low 1 High Inorganic Growth Index ------> solutions. Some of the significant Total R&D headcount in autonomous technology acquisitions being Mobileye (by Intel) and Cruise (by GM) Tech Mafias Semiconductor Auto Start-ups OEMS Tier 1s 16

Note: 1 Inorganic Growth Index: Function of investment in AV and related technology areas through Acquisitions, Partnerships and investment through Corporate Venture Arms; 2 Technology maturity Index: Function of maturity of technology across the AV stack of components, sub-systems and full-stack autonomous systems required to develop AV capability Source : DRAUP 16 The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Acquisitions and investments: Semiconductor giants and OEMs have been leveraging collaborative AV innovations and acquiring highly mature solutions to develop AV capability

Total External investment spend to acquire AV capability

$22–23 Bn

SEMICONDUCTOR OEM TECH MAFIA

~$18 Bn ~$3 Bn ~$1 Bn

Top Acquisitions

17 Corporate VC Spend Acquisition Note: 1-Technology spend includes employee compensation and related expenses along with spend on software, platforms and hardware tools required to develop In-house capability; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 17 Acquisitions and investments: Automakers are thinking ahead and collaborating with Technology providers and disrupters to move beyond their traditional business segments

Computing & Vision based 3D Mapping/ Network, Sensors HMI/UI-UX Vehicle Control perception Localization Connectivity

Automakers

Tier-1s

Technology Suppliers Cloud based open Customized algorithms Connected car location platform; of computer vision, application to connect provides mapping, machine learning mobile to car and traffic data Software Provides full stack dashboard ADAS systems

Provides full stack ADAS system

Platforms Intel-Mobileye will provide computing platform, sensing & localization expertise

Bosch is co-innovating with Formed the Automotive Hardware/ Nvidia for the AI based Provides data processing, and Edge Computing software systems for its computing SoCs along with Consortium with Toyota Processors sensor technology Sensors and connectivity to boost creation of maps and ADAS technology

Services/ BMW and Ford have collaborated with Operators ride sharing giants such as Lyft and Uber respectively largely to mine vehicle driving data Mobility Services Microsoft, Valeo & Renault Nissan group partnered to Ericsson and Toyota have Network, Security leverage Azure cloud platform partnered for developing 5G and cloud customization for data security, infrastructure for enabling connectivity and privacy V2V, V2X communications18

01 Autonomous Vehicle Overview

02 Technology Spend Analysis This section provides an overview of :  Global and US AV adoption analysis 03 Autonomous Vehicle Adoption  Miles driven by top companies in California

04 Bay Area–Deep Dive

05 Top Companies Deep Dive

06 Partnership Opportunities

19 AV Adoption across globe: US states and several other nations are relieving the regulations around Autonomous Vehicle testing on public roads

Autonomous Vehicles regulations by State and Central government organisations  Michigan being a traditional automotive Michigan engineering hub became the first state to approve the latest autonomous technology Florida allowing automakers to test their autonomous prototypes on public roads California even without a driver.  Governments of UK, Japan and Germany Singapore are cautious about the safety of current autonomous technology. Thus they have taken proactive regulatory measures by Arizona allowing testing only in the presence of a driver. Netherlands  Governments in geographies such as China Germany, UK and other European countries are not able to develop a concrete regulatory framework for Germany testing and assessing autonomous driving because they face challenges in defining Japan ethical laws relating to responsibility in accidents caused by fully autonomous vehicles. UK  The Netherlands’ Council of Ministers Israel recently updated its bill to allow tests without a driver. Shanghai issued its first Regulations Legal for testing Legal for testing Legal for testing Legal for testing Semi autonomous self-driving license, allowing automakers to Passed prototype with prototype on public prototype without prototype on public fleet services test their AVs on public rods. driver roads with driver driver roads without driver allowed AV* Adoption Index LOW HIGH 20 AV*: Autonomous Vehicle Note: Autonomous Vehicle regulations have been verified from reports published by Department of Motor Vehicle, California and other state regulatory bodies in respective geographies; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 20 AV Miles Driven: Level 5 Autonomous Vehicles have Millions of Test Miles to complete before they can be Consumer Ready; Companies are investing in Simulation platforms

1 Autonomous Test Miles Driven In California (2018) • Waymo and GM seem to be way ahead of the competition when it comes to real Waymo Miles to Go world tests but are way behind the GM Cruise Industry Standard

Apple • On-Road testing is a very lengthy process that could take years to Aurora Industry complete. Hence, companies are shifting their focus towards simulated testing Zoox Standard which can simulate all aspects of the Uber 11 Billion Miles autonomous drive without posing any risk to pedestrians or other motorists Nuro (to reach required safety levels as per • OEMs are still figuring out the right Baidu industry consensus) balance of testing AVs in real world Number of Autonomous scenarios and simulated environments Pony.Ai vehicles on road in California • Companies like Tesla, Apple and BMW Renault… • GM Cruise: 163 rely mostly on simulated testing of AVs Waymo: 125 Drive.ai • • Apple:69 10 Million of • Companies like NVIDIA, Electrobit, NVIDIA Autonomous Cognata currently provide Simulation Test Miles solutions for AV testing Daimler since 2009 • Testing through simulations also gives the ability to test countless variations in Note: Ford, Lyft and Tesla are top players in AV but have disengaged from California DMV Autonomous Vehicles road conditions, scale and cost. 3 Million Miles of Autonomous Test Driving driven in 2018 Autonomous • Research done by RAND Corporation Test Miles suggests that autonomous vehicles need driven since to drive 11 billion miles in testing before 2016 being ready for consumers while the company with the highest autonomous AV Simulation miles, Waymo has only completed 7 Testing Providers million miles in 10 years. 21

Note: 1-The data retrieved from the website of California DMV. The data reflects the number of test miles covered by AVs in the state of California from December 2017 to December 2018. Source : DRAUP 21 AGENDA

01 Autonomous Vehicle Overview

02 Technology Spend Analysis

03 Autonomous Vehicle Adoption

04 Bay Area–Deep Dive

05 Top Companies Deep Dive

06 Partnership Opportunities

22 Bay Area Deep Dive: In Bay Area, Automakers have established AV innovation labs to collaborate with Tech Mafias and disrupters, and explore new AV enabled mobility solutions

Innovators Followers Emerging Players

6.5–7K 55% 30% 15% Bay Area

Core R&D team of ~1,000 engineers, located Autonomous Vehicle Engineering Google’s To invest $1 Bn in San Francisco over next in the Bay Area, is largely focused on developing deep five years in AI and self-driving cars R&D Headcount in Bay Area learning software capability for advanced vehicle control and automation Opened a new Automated Driving Group in Silicon Valley and plans to invest $250 Mn on Tesla is building critical ADAS systems in-house and self-driving tech via its Intel Capital investment leveraging partner network with Nvidia and Bosch for arm. Intel also has 3 other autonomous R&D autonomous hardware capabilities. labs in Arizona, Germany and Oregon.

VW works in partnership with Stanford Invested $14 Mn on the new expanded R&D facility University for autonomous driving. Its in California and plans to add 1,100 workers to it’s research lab -Volkswagen Automotive new acquisition team at Cruise Automation Innovation Lab is located within the Stanford University campus Invested $1 Billion in AI startup Argo AI; Developed aDRIVE gaming environment for Acquired HERE maps for 3D mapping autonomous test driving technology

Uber poached around 50 researchers and engineers Bosch has an autonomous driving solutions from Carnegie Melon University’s Robotics Institute center in Palo Alto. It partnered with Daimler to to build its autonomous capability launch automated valet parking system 23

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned players. List of emerging players non-exhaustive; The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 23 AGENDA

01 Autonomous Vehicle Overview

This section provides an overview of : 02 Technology Spend Analysis  Top leaders in the AV Ecosystem  Analysis of top leaders by: o Positioning Strategy 03 Autonomous Vehicle Adoption o EV focus for AV o Commercialization Roadmap o Capability and prime acquisitions 04 Bay Area–Deep Dive across segments  Deep Dive analysiss of top leaders: o Waymo 05 Top Companies Deep Dive o GM o Ford o Uber 06 Partnership Opportunities o Tesla

24 AV Positioning Strategy: OEMs like GM, Ford and tesla are trying to master each level of automation whereas, Waymo and Uber are working towards level 5 leadership

Deploy Fleet of self- driving Bolt EVs for Waymo & Uber skipped ride-hailing service in US by 2019 Level 5 semi-autonomous levels • GM realised that to attain Full Automation to focus on level 5 technology leadership in the integration with OEMs industry EV focus is not Level 4 Capitalize Level 5 enough. High Automation capabilities to integrate level 3 SuperCruise in Level 3 Cadillac • Failing to build AV expertise Conditional will create a technology Automation Level 2 dependence in future Partial Launched Semi- towards giants like Waymo autonomous Automation Cadillac CT6 (Google), Uber etc. equipped with self- Level 1 driving system Driver Assistance ‘SuperCruise’ • GM’s out-of-the-box AV efforts are evident in their Level 0 deployment strategy trying No Automation to monetize each level of 2010 2012 2014 2016 2018 2020 2022 Autonomy

GM Tesla Ford Waymo Uber

Product focus Integration focus 25

Note: The infographic above shows analysis done on specific companies. There are several other companies working towards Autonomous Vehicles The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 25 AV focus in EV: Electric Autonomous Cars such as Hybrid, Plug-ins and Plug-in hybrids autonomous cars to cover the urban transportation landscape by the next decade

2016 2018 2020 2022 2025  58% of autonomous light-duty vehicle models are currently built over an electric powertrain Level 3 integrated while a further 21% utilize a Deploy Level 5 Bolt Cadillacs hybrid powertrain, according to a EV Fleets testimony submitted at the House Level 2 integration with Cadillac CT6 Partner with Energy & Commerce Committee. hybrid OEMs to deploy AV technology.  Top drivers for Electric Test Level 5 in Autonomous Vehicle adoption : Jaguar I-Pace  Regulatory restrictions Partner with OEMs relating to gas-mileage to deploy AV requirements. technology Test Level 5 in Activate Level5 in  Electric cars are easier for Chrysler Pacifica all models computers to drive due to through over-the- Deploy self-driving fewer moving parts and air updates (OTA) with a fleet of ford low maintenance. Test Self-driving fusion hybrids  Wireless charging with fleet of Volvo XC90 integrates seamlessly with autonomy Integrated level 5 capable hardware in Tesla Lineup  Self driving cars to populate urban Test self-driving Testing self-driving areas first due to better availability with a fleet of ford with a fleet of ford fusion of charging stations. The US fusion hybrids Integrated level 2 in tesla Department of Energy lists around model S & X 48,000 such charging stations across America.

Note: The infographic above shows analysis done on specific companies. There are several other companies working towards the automation of electric vehicles. The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 26 AV Commercialization Roadmap: Companies like Tesla are planning to offer OTA updates that will transform existing models towards self-driving capabilities

2016 2018 2020 2023 2025

Launch Level 5 Level-2 “SuperCruise” Deploy Fleet of Level-5 Bolt EVS and then establish Level 3 dominance Integrate advanced “SuperCruise” in GM Lineup • Waymo has driven nearly 7 million miles and is leading the competition of AV Focus on Level 5 Deploy Fleet of robot axis to establish AV Achieve 7.0 million test miles to prove AV domination integration Partner with automakers to deploy full stack AV solution GM is planning to prove leadership • technology expertise by

Focus on Level 5 Deploy Fleet of AV ride-hailing services deploying Level-5 ride- to establish AV Test phase: Uber’s Self-driving ford fusion and hailing service and by Volvo XC90 integration Partner with OEMs to deploy full stack AV integrating the expertise in leadership SuperCruise System to

Embed vehicles Level-2 Auto-Pilot Model S & X achieve Level-3 with AV Rollout Level-5 in Autopilot-2 embedded capabilities and Level-5 hardware Embedded in to vehicles through OTA updates deploy through “AutoPilot-2” Tesla lineup • Ford visions to have a level 5 OTA updates self-driving vehicles for ride- hailing and door delivery Skip Level-3 and Level-2 integration Rollout Level-5 Self-driving Ford fusion for ride- focus only on hailing and door delivery services services with level-5

Level 2 Level 3 Level 4 or above 27

Note: The infographic above shows analysis done on specific companies. There are several other companies working towards Autonomous Vehicles The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 Source : DRAUP 27 AV Capability Deep-dive: Mapping, Sensors, and automation control systems based start-ups are prime acquisition targets

Automation Full Stack AV Overall Stack Security Connectivity HD Mapping Sensor Fusion control system Solution Rating

Google Lumedyne Cybersecurity Android Auto Waze Technologies 510 Systems • GM’s acquire and invest AV strategy is a contrast to their Cruise OnStar Ushr Strobe traditional initiatives of Automation spending minimal on acquisitions

Ford Sync Civil Maps Velodyne Argo AI • Ford is one of the top leaders who have invested $1 bn in Argo.AI to deploy full Otto & Geometric autonomous vehicles for Uber Technologies Decarta Intelligence commercial purposes

Tesla is building in-house MapBox Bosch Intel • systems and services for AV capabilities

Acquisition Investment Partnership Inhouse

High Medium Low 28

Waymo LLC Key Autonomous vehicle Activities • Develop and test high performance LIDAR systems Location • Develop insightful tests that span the range of radar integration stages, including individual snapshot evaluation, fully integrated on self-driving Mountain View, vehicles, and fleet wide data mining California • Design and execute LIDAR field measurements and structured tests • Build motion planning and decision-making systems for the self-driving vehicles, ensuring that the behavior of our vehicles is safe, smooth, and Autonomous predictable to other road users Headcount • Building backend infrastructure for storing and processing many forms of map data 250-300 • Research new machine learning problems, models and algorithms • Design and manufacture of LiDAR systems Center R&D Spend • Develop car’s computer vision system that processes billions of pixels per second with very low latency $ 1.1 Bn • Develop autonomous vehicle system including optical modelling, camera hardware design, image quality, ISP pipeline, deep nets for detection and classification, and high level perception evaluation Center Level HQ & Hub Key Profiles

• Hardware Design Engineer Key Influencers • Hardware Engineer, LIDAR Validation John Krafcik • Audio Systems Engineer CEO • Robotics Software Engineer, Behavior Prediction • Software Engineer, Machine Learning Infrastructure Dmitri Dolgov CTO, VP Engineering • Software Engineer, Mapping • Software Engineer, Computer vision system Daniel Chu Director of Product

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned player. List of AV activities and profiles are non-exhaustive; Source : DRAUP 29 Global Footprint data curated by DRAUP and updated in April, 2019 GM: Engineering Center Deep Dive

GM Warren Key Autonomous vehicle Activities Cruise Key Autonomous vehicle Activities Technical Centre • Development and integration of analytical Automation • Computer vision and LIDAR-based solutions for algorithms and tools for autonomous vehicles. robotic perception Location Location • Development of Simulation platform for testing • Design, implementation and support of network Warren, Michigan and simulating autonomous cars San Francisco Bay monitoring and alerting systems Area Greater Detroit Area • Autonomous system integration with hardware • System and sub-system level requirements for and software redundancy, fault-tolerant focus perception and localization software Autonomous Driver modeling/machine learning • System and subsystem level validation planning Autonomous • Headcount development/integration and execution Headcount 1,300-1,400 2100-2200 • Functional safety, hazard analysis, risk • Safety analysis and gaps coverage assessment • Drawing and semantic annotation of road maps Total Center Spend Total Center Spend • Inspecting map labeling to ensure compliance for $ 728 Mn $ 1 Bn organizational standards

Center Level Center Level Key Profiles Key Profiles HQ & Hub Hub Autonomous Driving Software Engineer • • Autonomous Driving Software Engineer Key Influencers • Autonomous Driving Controls Engineer Key Influencers • GIS Mapping Technician Autonomous Vehicle System Safety Engineer Dan Ammann • Kyle Vogt • Autonomous Security Engineer President Autonomous Validation Engineer CEO • • Self-driving systems Engineer • Autonomous Performance Engineer Computer Vision Engineer Pamela Fletcher Daniel Kan • VP, Autonomous & EV Algorithm Design and Development Engineer • COO • Network Engineer Aaron Sullivan Engineering Manager Autonomous system

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned player. List of AV activities and profiles are non-exhaustive; Source : DRAUP 30 Global Footprint data curated by DRAUP and updated in April, 2019 Ford: Engineering Center Deep Dive

Ford Autonomous Key Autonomous vehicle Activities Vehicle LLC • Development and design of autonomous vehicle sensing components

Location • Architecturall design, execution and development of infotainment platform Dearborn, Michigan Greater Detroit Area • Provide quality assurance for both hardware and software components

Autonomous • Conduct performance design verification tests on prototype vehicles and constituent systems\ Headcount 1350-1400 • Write production quality code to deploy as Transport-as-a-Service solutions • Develop Remote sensing technologies Center Spend $ 900 Mn

Center Level Key Profiles HQ & Hub • Autonomous Vehicle Embedded Platform Software Architect Key Influencers • AV Sensor and Module D&R Engineer Sherif Marakby • AV - Systems Validation Engineer CEO, Ford Autonomous Vehicles LLC • AV- Software Engineer Robert Walker • Autonomous Vehicle Product Innovation Engineer AV Product & Experience Design Chief • Advanced Driver Assistance Systems and Controls Testing and Development Automated Driving Feature Development Engineer Chris Brewer • Chief Engineer, Autonomous Vehicles

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned player. List of AV activities and profiles are non-exhaustive; Source : DRAUP 31 Global Footprint data curated by DRAUP and updated in April, 2019 Uber: Engineering Center Deep Dive

Advanced Technology Key Autonomous vehicle Activities Advanced Technology Key Autonomous vehicle Activities Group Center • Design, implement and optimize novel Group Center • Create android applications for self-driving algorithms that run at extremely low latency on systems Location autonomous vehicles Location • Manage Autonomous Vehicle Integration program Greater Pittsburgh San Francisco Bay • Define, develop, implement and maintain the delivery, including milestones, prototype builds, Area manufacturing requirements and test Area and launch specifications • Establish process to manage changes for all Autonomous Carry out root/cause analysis component builds and vehicle builds Autonomous • Headcount Manage Autonomous Vehicle Integration • Interface with Vehicle OEMs and Tier1 suppliers to Headcount • 500-700 250-300 program delivery align technology and vehicle delivery Center Spend • Work with lidar sensor firmware and low level Center Spend $ 77 Mn signal processing $ 232 Mn

Center Level Center Level Key Profiles Key Profiles HQ & Hub Hub AI Research Scientist • • Embedded Software Engineer Key Influencers • ATG Manufacturing Test Engineer Key Influencers • Autonomous Vehicles-Embedded Eric Hanson • AV Maps Quality Analyst Carl Wellington Verification and Test Engineering Head of Product, Advanced Autonomous Vehicle Program Manager Director, Self Driving Cars Technologies Group • • Autonomous Vehicle Program Manager Sameer K Android Engineer, Self-Driving Experience • Jon Thomason • Computer Vision Engineer Supply Chain Director, VP, Software Advanced Technologies Engineering • Backend Engineer, Self-Driving Group Brian Zajac Steven Choi Head, Systems Product & Strategy Engineering & Testing

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned player. List of AV activities and profiles are non-exhaustive; Source : DRAUP 32 Global Footprint data curated by DRAUP and updated in April, 2019 Tesla: Engineering Center Deep Dive

Key Autonomous vehicle Activities

Tesla • Work on the Camera software pipeline running on the target product platform to deliver high resolution images at high framerate to a range of consuming devices (CPU, GPU, hardware compressors and image processors) Location Palo Alto, California • Optimize and integrate embedded code to introduce new features and capabilities to Tesla’s vehicles.

• Develop state-of-the-art algorithms in multi-sensor fusion, visual-inertial odometry, GPS, IMU and radar processing, intrinsic/extrinsic camera Autonomous calibration, structure from motion, etc. Headcount 300-400 • Develop software platform and tools for AI algorithms in self driving cars. Define system reliability and robustness requirements for the autopilot ECU Center Spend • $ 392 Mn

Center Level Key Profiles HQ & Hub • Computer Vision Scientist/Engineer, Autopilot Key Influencers • Firmware Engineer, Autopilot Andrej Karpathy • Autopilot Systems Design/Functional Safety Engineer Senior Director of Artificial Intelligence • Autopilot Software Engineer, Computer Vision and AI Neeraj Parik Autopilot - AI Technical Lead Architecture and Design • (Autopilot Hardware) • Architect, IoT Technology Lead Mitchell Heschke Sr. Product Design Engineer- Autopilot

Note: Includes recent R&D initiatives and collaboration announcements/activities of the above mentioned player. List of AV activities and profiles are non-exhaustive; Source : DRAUP 33 Global Footprint data curated by DRAUP and updated in April, 2019 AGENDA

01 Autonomous Vehicle Overview

02 Technology Spend Analysis

03 Autonomous Vehicle Adoption

04 Bay Area–Deep Dive

05 Top Companies Deep Dive This section provides an overview of :  Partnership opportunities across AV areas  Outsourcing intensity across AV areas 06 Partnership Opportunities

34 Partnership Opportunities: High partnership opportunities in System Engineering, System integration & Feature development.

Outsourcing Intensity Prototype Testing & System Safety System Modelling Validation System Engineering & Validation of autonomous System Architecture alignment Functional Safety System validation features against diverse road to the autonomous vehicle’s including performance scenarios Full Full Stack mission validation

ADAS Algorithms Computer Vision/ML Virtual Environment Simulation & Testing Feature Algorithm development • Image Processing and Development and calibration, validation, Machine Vision • Virtual environment and functional safety. • Traffic Incident development and data readiness and ML collection

ECU platforms Maps & Navigation Vehicle Control Systems Data Analytics & Platforms Cyber Security System Integration • Integration software GPS/INS-based vehicle • Driver Assist Systems (Middleware) platform state estimation, 3D • Emergency Breaking • OTA Software • Platform testing mapping and • Steering Control Management localisation

Physical Design & FPGA/ SoC Testing

SoC Testing, Physical • SoC Verification Design • SoC Validation • Burn-in stress testing

Sensor Testing & Validation Sensor Perception Sensor Design Software • RADAR • Night time Pedestrian detection • Camera Module Reference Design Sensor Fusion • LIDAR • Distance and Angle Estimation • Radar Module Design

• Ultrasonic • Pedestrian, cyclist detection • Multi-Sensor Hub Reference Components Components & Hardware • Camera • Path planning & object tracking Design Note: The above analysis is based on the outsourcing done by the OEMs, Tier-1 Suppliers, Tech Giants and Start-ups in the AV ecosystem. There are several other companies Source : DRAUP 35 working towards Autonomous Vehicles. The above analysis is based on the DRAUP’s proprietary engineering database and insights from industry stakeholders, updated as on April 2019 www.draup.com 36

Source : DRAUP 36