Has the death knell for the commercialization of self-driving taxis sounded for startups? Behind Drive.ai's "sale"

Latest update time：2019-03-05

Reads：

▲Click above Leifeng.com Follow

Text | Yu Shengyue

Report from Leiphone.com (leiphone-sz)

Drive.ai, which is backed by Andrew Ng and has a team of 170 people and has already started operating self-driving taxi services, is also on the road to "selling itself".

According to the well-known technology media The Information, Jefferies Bank has obtained the contract with Drive.ai. As an advisor to Drive.ai, the investment bank will pass on the news that Drive.ai is seeking to sell to other autonomous driving companies in recent weeks.

If the transaction goes smoothly, Drive.ai may become the first Chinese self-driving startup to successfully sell itself. Otherwise, the only option left for Drive.ai is to continue to raise funds, increase investment, and survive in the torrent of self-driving.

Drive.ai has completed two rounds of financing, totaling $77 million. In June 2017, Drive.ai received $50 million in Series B financing; three months later, the company received another $15 million in venture capital. At that time, Drive.ai's valuation had reached $200 million, and its list of investors was very long, with support from eight companies including Northern Light Venture Capital, GGV Capital, NEA Enshuo Investment, and Grab.

Why is Drive.ai, which has achieved commercialization, going to "sell itself"?

Drive.ai was founded in 2016 and was born out of Stanford University's Artificial Intelligence Laboratory. The person who endorsed this small company was Carol Reiley, the president of Drive.ai and the wife of AI expert Andrew Ng. Later, Andrew Ng also joined the Drive.ai board of directors as a director.

Recently, there are reports that Andrew Ng and his wife Carol Reiley have withdrawn from the company's operations. The company's operations are run by several of Andrew Ng's students. There are also internal conflicts among these people, and some core members have sought opportunities to return to China. In early February, co-founder Joel Pazhayampallil announced his departure from the company.

Internal turmoil was also identified by the industry as one of the reasons why Drive.ai sought to "sell itself".

In terms of commercialization, Drive.ai has been trying to build its own autonomous driving platform and create cost-effective autonomous driving software using deep learning and artificial intelligence. Drive.ai claims that their "deep learning first" approach allows autonomous vehicles to quickly understand new driving scenarios and routes, saving a lot of tedious engineering work, thus achieving the goal of rapid and large-scale deployment. Last year, Drive.ai completed the commercialization of autonomous driving:

In July 2018, Drive.ai's first batch of 3-4 (4 during peak hours) self-driving vans began pilot service between the Hall Park campus and The Star business district in Frisco, Texas, starting a 6-month operation, providing free last-mile travel services for tens of thousands of employees in the park;

Three months later, Drive.ai announced that it would officially launch an autonomous driving taxi service in Arlington, Texas, to residents, employees and tourists. This is also the first truly public autonomous driving ride service in Texas.

The small company completed two commercial deployments within six months, a speed that was once astonishing.

The difference between the commercial autonomous driving services provided by Drive.ai and Waymo in Phoenix is that the latter has started charging passengers, while the former is still in the free service stage, mainly conducting commercial cooperation with the government, commercial parks or where the projects are located.

Although Drive.ai's commercial autonomous driving service has received government funding: the city of Arlington provides Drive.ai with some funding, and the U.S. federal government will also allocate $343,000 to support it, a few hundred thousand dollars is still a drop in the bucket for capital-intensive autonomous driving.

Leifeng.com New Intelligent Driving learned that Drive.ai currently has nearly 20 self-driving vans in Texas. After the news of the "sale" came out, it was also questionable whether these services can continue.

In terms of autonomous driving technology, many startups such as Pony.ai and Drive.ai are following the footsteps of Waymo, the "big brother" of autonomous driving, aiming at high-level L4/L5 autonomous driving and developing autonomous driving fleet operation services. However, Drive.ai's technical capabilities are not outstanding.

Drive.ai’s road test project was officially launched in May last year, which is indeed a step behind its competitors. Many well-funded leading players had already put their test cars on the streets a few years ago.

The 2018 Autonomous Driving Takeover Report released by the California Department of Motor Vehicles shows that Chinese-founded autonomous driving companies Pony.ai and WeRide are both ranked ahead of Drive.ai. Drive.ai ranks 12th, requiring manual takeover once every 134.3 kilometers on average. Pony.ai ranks 5th, requiring manual takeover once every 1,600 kilometers on average.

Is the commercial operation of self-driving taxis a panacea for startups?

As an industry with scarce talent and capital intensiveness, for start-ups, lack of funds means it is difficult to find suitable R&D personnel, difficult to form a fleet to complete experiments and tests, and even more difficult to survive.

Leifeng.com's New Intelligent Driving learned that the domestic autonomous driving fleets formed by Pony.ai and WeRide are still limited to operating within a limited range, with safety drivers in the cars, and have not achieved true commercial profitability of autonomous driving.

Waymo, the "big brother" of autonomous driving, once confidently "kicked" the safety driver out of the driver's seat, but soon after, it invited the safety driver back again. Therefore, the inability to solve the driver cost problem is still a major factor hindering the real profitability of autonomous driving. The cost of safety drivers is even higher than the cost of drivers.

On the other hand, Waymo's commercial operation of autonomous driving is still limited to Phoenix. Large-scale fleet building and commercial operation of autonomous taxis on a larger scale face technical and cost pressures, which is not easy for the technology giant Waymo, let alone a startup with weaker technical capabilities and insufficient funds.

Senior investors in the industry said that although capital has poured into the autonomous driving industry, the current returns are minimal. For those companies that have inherent technical constraints, insufficient self-sustaining capabilities, unformed business models, and are unlikely to see profits in the short term, they will face great risks in the current capital market environment.

The industry's expectations for the large-scale implementation of L4 and higher levels of autonomous driving have been continuously delayed. For autonomous driving startups that are far smaller than Waymo, the biggest problem is how to survive in the next decade.

Low-speed autonomous driving in limited scenarios is favored

At the end of last year, Waymo CEO John Krafcik personally admitted that autonomous driving technology is difficult and that the ultimate stage of L5 autonomous driving is a somewhat unreasonable expectation and even unnecessary.

Therefore, before the L4/L5 high-speed autonomous driving technology for all road conditions matures, many companies choose to rely on low-speed autonomous driving (such as automatic valet parking, last-mile logistics, and sanitation) to achieve rapid implementation and help the company "survive."

Low-speed autonomous driving in limited scenarios, such as autonomous driving logistics and automatic parking, has already been commercialized and is favored by capital.

In the spring of 2019, Nuro, which focuses on driverless delivery services, completed a record-breaking $940 million in financing from SoftBank Vision Fund. Nuro's driverless delivery service has now been commercialized and is providing autonomous driving delivery services to the people of Scottsdale. This is also the world's first truly driverless car delivery project for the public. Because Nuro's driverless delivery vehicle R1 running on public roads will officially bid farewell to drivers and passengers, and only transport goods.

Leifeng.com New Intelligent Driving learned that TuSimple, an autonomous driving company focusing on logistics scenarios, also recently announced the completion of a $95 million Series D financing. After realizing the growing demand for automatic valet parking last year, traditional OEMs, Tier 1 suppliers and new car-making forces at home and abroad began to take action and shouted slogans to make future plans for the mass production of automatic valet parking.

The public can also see that there are only a few companies that are profitable through commercialization of autonomous driving, and most companies are still in the "capital input" state. At this stage, starting from limited low-speed scenarios and entering the high-speed autonomous driving market when the time is right in the future has become the business route of most autonomous driving startups.

If you want to get a limited-time free ticket to the summit, scan the QR code in the picture above to register. After passing the review, you can get the ticket to the Leifeng.com New Intelligent Driving 2019 AI+Smart Car Innovation Summit .

In addition, the "AI+Smart Car Innovation Summit" WeChat community has been opened for practitioners, product and technology developers in the smart car industry. You are welcome to add WeChat ID: eriseds to join the group.

- END -

◆ ◆ ◆