How China’s first autonomous driving unicorn Momenta hunts for data

by

in

Cao Xudong turned on the side of the street at panties and a black T-shirt printed with the term “Momenta,” the name of his startup.

Before founding the firm — that last year topped $1 billion at valuation to become China’s first autonomous driving “unicorn” — he’d led an undercover life, but he was convinced that autonomous driving are the true big thing.

Cao isn’t only going to the moonshot of completely autonomous vehicles, which he says might be 20 years away. Rather, he’so taking a two-legged approach of selling semi-automated applications while investing in search for next-gen self-driving technology.

Cao, pronounced ‘tsao’, was chasing his Ph.D. in engineering mechanics as soon as an opportunity came to work in Microsoft’s basic research arm in Asia, putatively that the “West Point” for China’s earliest creation of artificial intelligence experts. He held out there for at least four years before quitting to put his hands on something more practical: a startup.

“Academic research for AI was becoming quite old at the moment,” said now 33-year-old Cao in an interview with TechCrunch, reflecting on his decision to stop Microsoft. “But the business that puts AI into application had only started. I believed the industrial wave could be much more extensive and intense compared to the academic wave that lasted in 2012 to 2015. ”

In 2015, Cao combined SenseTime, now the world’s highest-valued AI startup, thanks in part into this profitable face-recognition technology it sells to this authorities. Throughout his 17-month stint,” Cao built the company’s research division from zero team into a 100-people strong team.

Before long, Cao found himself craving for a new experience again. The creator said that he doesn’t care about the result as much as the opportunity to “do something. ” That trend was evident through his time in the esteemed Tsinghua University, at which he was a part of the outside club. He wasn’t especially drawn to hiking, ” he said, but the chance to adopt challenges and be with similarly resilient, so adventurous people was luring enough.

And when creating driverless vehicles will let him leave a mark on the planet, he’s in for that.

Make the personal computer, not the car

Cao walked me to a car equipped with the cameras and radars you may spot on a autonomous vehicle, with unseen computer codes installed in the trunk. We jumped in. Our driver picked a route from the high-definition map that Momenta had constructed, and as soon as we approached the highway, the autonomous mode switched on alone. The sensors then began feeding real-time information about the environment into the map, with which the computer may make decisions on the street.

momenta

Momenta staff installing sensors into a testing car. / Photo: Momenta

Momenta won’t create hardware or cars, Cao ensured. Instead, it provides cars autonomous features by making their brains, or deep-learning capacities. It’s in effect a so-called Tier two provider, comparable to Intel’therefore Mobileye, that sells to Tier 1 providers who actually create the automotive components. Additionally, it sells directly to original equipment manufacturers (OMEs) that design cars, purchase parts from suppliers and assemble the last product. Under both circumstances, Momenta functions with clients to define the last piece of applications.

Momenta considers this asset-light approach would let it develop state-of-the-art driving tech. By selling applications to parts and car manufacturers, it not only brings in income but also sources mountains of information, such as how and when humans intervene, to educate its codes at comparatively lower expenses.

The business declined to discuss who its clients are said they include high carmakers and Tier 1 providers in China and overseas. There won’t be many of them because a “venture ” in the automobile sector demands profound, resource-intensive cooperation, so less is believed to be . What we do understand is how Momenta counts Daimler AG for a backer. It’s the very first Chinese startup which the Mercedes-Benz parent had invested in, though Cao wouldn’t disclose if Daimler is a client.

“Say you operate 10,000 autonomous cars to reap information. This might easily cost you more $1 billion a year. 100,000 cars could cost $10 billion, and it is a frightful amount for any technology giant,” Cao said. “If you wish to get seas of information that have a meaningful reach, you have to construct a product to the mass market. ”

Highway Pilot, the semi-autonomous solution that was controlling our car, is Momenta’s first mass-produced software. More will launch in the forthcoming seasons, such as a totally autonomous parking alternative and also a self-driving robotaxi bundle for urban usage.

In the very long run, the startup said it aims to tackle inefficiencies in China’s $44 billion logistics market. People hear about warehousing robots constructed by Alibaba and JD.com, but total, China remains at the lower end of logistics performance. In 2018, logistics prices accounted for nearly 15 per cent of national gross domestic product. In precisely the exact same season, the World Bank rated China 26th in its logistics performance index, a global standard for efficiency in the business.

momenta

Cao Xudong, co-founder and CEO of Momenta / Photo: Momenta

Cao, an unassuming CEO, increased his voice as explained the company’s two-legged plan. The twin approach forms a “closed loop, so ” a word that Cao repeatedly summoned to talk about the company’s competitive advantage. Instead of picking between the existence and future, as Waymo does with Level 4 — that a designation awarded to cars that may operate under fundamental situations without human intervention — and Tesla with half-autonomous forcing , Momenta functions on both. It uses revenue-generating businesses for example Highway Pilot to finance research in robotaxis, and the sensor information collected from real life scenarios to feed models in the lab. Results in the lab, subsequently, could soup up what gets set up on public roads.

Human or machine

Throughout the 40-minute ride at midday visitors, our car managed to change lanes, merge in traffic, generate space from reckless drivers alone except for one short moment. Toward the end of the trip, our driver forced the lever to activate a lane shift because we approached a car dangerously parked at the middle of the exit ramp. Momenta titles this “interactive lane shift,” that it claims is intended to be part of its automated system and by its rigorous definition isn’t a person “intervention”.

“Human-car interaction may continue to rule for quite a while, possibly for another 20 years,” Cao noted, adding the installation brings security to another level since the car knows exactly what the driver is doing through its inner-cabin cameras.

“For example, in the event the driver is looking down in their phone, the [Momenta] program will alert them to listen,” he explained.

I wasn’t allowed to picture during the journey, and here’s some footage from Momenta to provide a sneak peek of its highway solution.

Human beings have been farther along the autonomous spectrum than most people think. Cao, like a lot of additional AI scientists, considers robots will gradually take the wheel over. Alphabet-owned Waymo has been operating robotaxis at Arizona for many months now, and smaller startups such as Drive.ai are also offering a comparable service at Texas.

Despite all the hype and flourish in the business, there stays unanswered concerns around passenger security, regulatory schema and a slew of other problems for the high-tech technology. Uber’s deadly self-driving crash last year postponed the business ’s potential jobs and motivated a public backlash. As a Shanghai-based venture capitalist recently indicated to me“I don’t believe humanity is ready for self-driving. ”

The biggest problem of this business, he contended, isn’t tech-related but societal. “Self-driving poses challenges to society’s legal strategy, civilization, justice and ethics. ”

Cao is well aware of the contention. He acknowledged that as a firm with the ability to steer prospective cars, Momenta must “endure a lot of responsibility for security. ” As such, he needed all executives at the company to ride a particular number of autonomous miles if there’s no loophole in the machine, the supervisors will probably stumble across it before the clients do.

“With this policy in place, the direction will pay serious attention to system security,” Cao asserted.

Momenta

Momenta’s new headquarters at Suzhou, China / Photo: Momenta

Regarding actually designing the applications to be dependable and to trace responsibility, Momenta appoints an “architect of program development and research,” who basically is responsible for analyzing the black box of autonomous driving calculations. A profound learning version must be “explainable,” said Cao, which will be key to discovering what goes wrong: Is it that the sensor, the personal computer, or even the navigation app that’s working?

Going ahead, Cao said that the company is in no rush to create a profit as it’s still spending heavily on R&D, but he assured that margins of this software it sells “are large. ” The startup can be blessed with sizable fundingsthat Cao’s resume surely helped pull, and thus did his additional co-founders Ren Shaoqing and Xia Yan, who were also alumni of both Microsoft Research Asia.

As of last October, Momenta had raised at least $200 million from big-name investors such as Daimler, Cathy Capital, GGV Capital, Kai-Fu Lee’s Sinovation Ventures, Lei Jun’s Shunwei Capital, Bluelake Capital, electric vehicle manufacturer NIO’s investment , WeChat operator Tencent and the authorities of Suzhou, that will house Momenta’s new 4,000 sq-meter headquarters right next to the city’s high-speed route channel.

After having a bullet train speeds past Suzhou, passengers can see from their windows Momenta’s familiar M-shape building, and this, in the years ahead, could turn into a new landmark of this historical city in eastern China.

Update (June 14, 2019): The post was updated to fix investors’ titles and clarify that the driver made a turn by pushing the lever, and not steering the wheel.

Buy Tickets for every event – Sports, Concerts, Festivals and more buy tickets

Discover more from Teslas Only

Subscribe now to keep reading and get access to the full archive.

Continue reading