Boston Consulting Group estimates that the annual value of the autonomous vehicle market will reach USD 77 billion by 2035. As of January 11, 2018, 50 companies have obtained autonomous car testing permits from the California Department of Motor Vehicles. Major tech companies, automobile manufacturers, automobile parts suppliers, and hundreds of startups have joined this sizeable tech boom.
VC firm Comet Labs created a diagram (see below) to demonstrate the diversity and complexity of the players involved. The autonomous driving industry is changing rapidly, and investment, mergers, collaborations, and failures happen every day. Taylor Stewart from Comet Labs refers to the industry as “a spider web all over.”
This article lists 10 companies which are representative of the leaders, followers, and challengers in this constantly degenerating and regenerating industry.
Waymo is the an Alphabet company aimed at fully self-driving vehicles that has been testing its vehicles in multiple locations including Google’s headquarters in Mountain View, California. Google launched the self-driving car project in 2009, and the project later became Waymo LLC in 2016. Waymo has collected more than 4 million miles of self-driving data, a distance that would take the average American driver nearly 300 years to complete. The Early Rider Program accepts early riders in the Phoenix metropolitan area to test its vehicles and share their thoughts.
“Fully self-driving vehicles will succeed in their promise and gain public acceptance only if they are safe,” says the 42-page Safety Report released in October 2017. Waymo defines “safety” in the 5 categories of behavioral safety, functional safety, crash safety, operational safety, and non-collision safety. All this is accomplished by incorporating a multi-layered sensor suite, computing and machine learning technology, backup systems for redundancy, as well as cybersecurity.
Policy makers have began introducing insurance rules for driverless cars. Insurance companies are likewise considering how to best respond to the threat driverless technology might pose to their businesses. Waymo teamed up with Insurance startup Trōv in December to cover riders in its driverless cars. The plan will be the world’s first trip protection for Waymo’s forthcoming ride-hailing service. California regulators proposed that manufacturers of autonomous vehicles should pay $5 million in liability insurance to provide coverage. However, Trōv’s insurance policies would only cover people inside driverless cars.
Alphabet’s recent fourth quarter earnings release shows the cost of running Waymo is high. Waymo still runs at a loss, but its losses are shrinking. The good news is that Fiat Chrysler said in January it will provide Waymo with thousands of Pacifica hybrid minivans for Waymo’s public ride-hailing service later this year. With the soon-to-come ride-hailing service, Waymo will likely scale its profit soon.
On the other hand, the ordinarily pioneering and outspoken Tesla has disappointed fans in its development of autonomous vehicle technology. Tesla already confidently sells vehicles with its ‘Fully Self-Driving (Level 5)’ feature, which is currently said to be disabled.
Tesla claims they will enable it through software updates as soon as regulators allow it. Evidently, it appears regulations have become a hurdle for Tesla.
Another area of discourse centers around Tesla’s refusal to use of LiDAR, which can be thought of as a radar that uses pulses of light instead of pulses of radio waves to probe the surrounding environment. LiDAR provides depth information, but its low resolution makes it hard to detect small and faraway objects without help from a normal and cheaper camera. Self-driving cars often use both cameras and LiDAR. The high cost of LiDAR delays the rollout of autonomous cars to some extent. Tesla only relies on cameras for its driving system, Autopilot. The debate around LiDAR will continue, as even academic institutes remain divided as to whether it’s a crucial component of autonomous driving vehicles.
Things have been rough for Tesla, in terms of autonomous driving. The company tested zero autonomous cars on California roads throughout the 2017 year. Last spring, a Tesla Model S driver was killed in a crash in Florida when the car’s autopilot system failed to identify a truck. Then, Tesla’s July breakup with Mobileye pushed the safety issue to its peak. Mobileye is the supplier to Autopilot. Mobileye condemned Tesla’s safety oversights in using Autopilot on a hands-free basis. However, Tesla maintains that Mobileye’s jealousy towards Tesla developing its own system (Autopilot) was the primary cause for the breakup. It may be a two-sided story, but the divorce was described as ugly by many media outlets including Bloomberg.
The consulting and research firm Navigant has scored autonomous driving companies since 2015. Tesla was the lowest ranked ‘challenger’ in the assessment due to the company’s poor performance in both strategy and execution of autonomous technology. Notably, Elon Musk, in a TED Talk last year, over-promised that a fully automated Tesla would be possible by 2019. However, NVIDIA, the company’s chip provider, has expressed doubt that the computing hardware it sells to Tesla is capable of reliably supporting full automation. As a result, Tesla is now working on its own chip for its self-driving software in partnership with AMD.
Compared to Waymo’s highly visible and aggressive development, Apple’s self-driving technology is not well known to the public. Apple isn’t readily competing with Waymo or Tesla in building its own vehicle but is instead focusing more on the software side of autonomous technology, such as commuting, processing, and algorithm systems. Apple is interested in automated systems in general, but has never specified application of its technology in vehicles. Apple CEO Tim Cook said the “autonomous system is the mother of all AI projects” at an interview with Bloomberg, suggesting that he considers autonomy a core technology for many applications including autonomous vehicles. Cook is very excited because autonomy, electrification, and ridesharing are happening within the same time frame.
Apple has made significant software advancements in recognizing small objects such as pedestrians and cyclists using LiDAR alone. The company’s experiments were based on computer simulations, however, and did not involve road tests.
4. Mobileye x Intel
Mobileye develops vision-based driver-assistance systems providing warnings for collision prevention and mitigation. On the company’s website, Mobileye claims that “25+ global automakers rely on Mobileye technology, 15 million+ vehicles are equipped with Mobileye technology, and 13 automakers are working with Mobileye to enable autonomous driving.” Mobileye emphasizes three pillars that make autonomous driving a reality: sensing, mapping, and driving policies. In terms of sensing, Mobileye believes the combination of LiDAR, camera, and radars together make for a reliable sensing system.
Mobileye was founded in Israeli and acquired by Intel in 2017. Intel announced more than USD 250 million in investment will be allocated towards autonomous driving technology. Later, Intel spent USD 15 billion to purchase Mobileye, which has come to shape Intel’s position in the autonomous driving race. Intel’s vision is to combine its 30 years of experience in security and IT with Mobileye’s vision processing technology to create the complete package. Intel aims to deliver a product that amalgamates a vehicle with a communications and data center.
Intel established a partnership with Waymo in 2017. The Intel x Mobileye-based technologies for sensing, computing, and connectivity enable Waymo to make real-time decisions and achieve advanced processing power required for Level 4 and 5 autonomy.
NVIDIA is traditionally known as a GPU (graphics processing unit) maker. NVIDIA utilizes its strong research capabilities in visual processing and initiated its own involvement in the autonomous driving field. The NVIDIA DRIVE PX systems fuses data from cameras, LiDAR, radar, and ultrasonic sensors, allowing algorithms to accurately understand the surrounding environment around a car.
NVIDIA’s long list of partners covers more than 320 entities and is still growing. On the list, you’ll find traditional automakers like Volkswagen, Audi, and Mercedes Benz, as well as tech behemoths such as Tesla, Uber, and Baidu.
The automotive business accounted for 19% of NVIDIA’s total revenue from the Q3 2018 fiscal report. Though a relatively small portion of NVIDIA’s overall business, automotive-related revenue tripled from just 7% of the Q1 2018.
As a rule broker of the ride-hailing market, Uber was once the leader of unicorn firms. However, since early 2017, the ride-hailing giant has suffered a great deal of negative publicity, including sexual harassment in the office, then-CEO Travis Kalanick’s squabbles with Uber drivers, and increasing surge pricing during the taxi driver strike in response to President Trump’s refugee travel ban. Yet, all these controversies are incomparable to the contentious Uber-Waymo trial.
Uber expects taking drivers out of the equation to gain 100% of the fare from the vehicles running 7×24. Kalanick met Anthony Levandowski while Levandowski worked for Google’s autonomous driving unit. Uber wanted to hire Levandowski, while Levandowski wanted to start his own company. Levandowski cofounded Ottomotto with Lior Ron. Only 6 months later, the company was acquired by Uber for a reported USD 590 million in 2016. “I tried to come up with a situation where he could feel like he started a company and I could feel like I hired him,” Kalanick testified in February 6’s federal trial.
Uber’s approach might be described as win-at-all-costs to catch up its autonomous driving technology with its peers. But it turned into a lawsuit in which Waymo contends that Uber stole its intellectual property. Waymo showed much evidence to support its claim.
On February 9, Uber and Waymo reached a settlement. Uber agreed not to use any of Waymo’s confidential information in its hardware or software. Uber gives Waymo 0.34% of its equity. Uber did not lose the war entirely. Uber earned its respectability back. Naturally, it is critically important for Uber to wash away the tarnish of bad publicity and resurrect its brand image. Waymo on the other hand can potentially benefit from Uber’s future successes. The decision ultimately may allow opportunity for Waymo and Uber to work collaboratively in the future.
7. Didi Chuxing
Didi Chuxing, which acquired Uber China in 2016, opened its AI labs in Silicon Valley in 2017. Didi hired Charlie Miller, the former head of Uber’s autonomous driving department. Miller will lead Didi’s safety and security department in and effort to make Didi’s autonomous vehicles resistant to external cyber attacks and threats.
Didi announced it will build an electric vehicle-sharing network with 12 automakers including a Renault-Nissan-Mitsubishi Alliance. “This cooperation fits with the Alliance’s expansion in vehicle electrification, autonomy, connectivity and new mobility services,” Ogi Redzic, senior vice president for Renault-Nissan-Mitsubishi, said in a statement.
Baidu’s strategy for autonomous driving is tied closely with its AI development. Its core self-driving platform is known as Apollo. Baidu opted to make Apollo an open source platform, allowing a degree of freedom for developers to improve it. The free resource include HD mapping, Autonomous Driving simulation engine, and deep learning algorithms. Apollo will help many Chinese local automakers, suppliers, and startups to design their autonomous driving vehicles and platforms. Additionally, it will likely further strengthen Baidu’s leading position in China.
Baidu invested heavily in AI after hiring AI expert Andrew Ng. Yet, Ng left Baidu in March 2017. Additionally, Jing Wang, Baidu’s former GM for the company’s Autonomous Driving Unit, also announced his own departure following Ng’s resignation. Baidu’s value on stock market took a USD 1.5 billion plunge after the senior scientists’ exit. Now, Baidu’s next move in autonomous driving technology remains unclear.
After leaving Baidu, Jing Wang founded Jingchi. Jingchi has a strong senior strategic and managerial team, including Tony Han, the former Chief Scientist of Baidu’s Autonomous Driving Unit, Qing Lu, the former CFO of Velodyne LiDAR, and Qingxiong Yang, the former Senior Director of the autonomous driving at Didi.
Jingchi plans to test 50 self-driving taxis in Anhui Province, China by March. During the test period, the taxi will have a “safety officer.” NVIDIA has joined a group of investors in investing a total of $52 million in Jingchi in September.
Like the Waymo-Uber trial, Baidu is suing Jing Wang for allegedly leaking and repurposing its closely guarded secrets in autonomous driving.
When talking about driving accidents, we naturally link them with long-distance driving fatigue. 30% of Chinese truck drivers experience driving fatigue. Due to increasing e-commerce delivery and an insufficient supply of young drivers, logistics systems in China are facing a huge truck driver shortage. Chinese startup TuSimple has received sizeable investment from Sina, and NVIDIA is focusing on Level 4 truck autonomy in which the operational design domain of a vehicle is expected to perform safety-critical driving functions and monitor roadway conditions for an entire trip.
TuSimple’s business model is similar to Levandowski’s Ottomotto. TuSimple acquires trucks then retrofits them with advanced technology. TuSimple’s trucks are equipped with a self-developed computer vision algorithm and paired with radars and 3D-HD mapping technology. Level 4 autonomy can significantly improve driver welfare and eliminate reckless driving.
Self-driving trucks could cut logistics costs by 40% in the US and 25% in China as they can travel longer durations of time without rest and save at least 10% on fuel costs.
Do you remember the beginning of this article? The autonomous driving field is changing so rapidly everyday. While writing this article, I had to make changes on a daily basis. No doubt there are battles yet to come in this war, but let’s hope the technology will ultimately make for safer roads and highways and create a reliable and robust mobility network.
(Top photo from Sohu.com)