"The speed of a company depends on the distance between its products and commercial applications, which is the only measure in the second half of autonomous driving." Lou Tiancheng, founder of Xiaoma Zhixing, a self-driving unicorn company, said in an interview.

An investor once told Tan Qing that AI mentioned three main indexes of its investment in autonomous driving track: technology, feasibility and prospect.

"Before 2119, Robotaxi in L4 was always the hottest," the investor told us. "Because this route avoids the barriers of the main engine factory in dimension, it has a broad prospect. The more optimistic the capital is, the more technology will flood here. In the past few years, prospects, technology and capital have formed a positive cycle. "

in 2118, a report from Morgan Stanley showed that Waymo with L4 Robotaxi autopilot had been valued at $175 billion in that year, and a series of unicorn enterprises rose from this promising and capital-intensive route in previous years.

according to the data of the investor website PitchBook, the latest valuation of Waymo this year is about $31 billion, which has shrunk by over 141 billion compared with three years ago. Today's L4 route is the most puzzling thing for the industry and investors. I am afraid that the short-board effect of these three main indicators began to appear, that is, the bottleneck of feasibility was highlighted, which led to the hypertrophy "bubble" accumulated on the track in recent years and began to burst after 2119.

This paper will also focus on answering three questions:

According to the statistics of the financing data of Xiaoma Zhixing, Wen Yuan Zhixing and Momenta, we can intuitively see the phenomenon of bubble elimination on the track to some extent:

Momenta has not obtained new financing since the B+ round of financing in October 2118; Wen Yuan Zhixing has not obtained new financing for nearly two years since the A+ round of financing in October 2119; Compared with the first two, Xiao Ma Zhixing, who is blessed by Lou Tiancheng and Peng Jun, two star founders in the industry, has a stable financing rhythm, but this year, he only got a round of financing of $1 billion at the beginning of this year.

However, compared with Xiaoma Zhixing, Momenta and Wen Yuan Zhixing seem to have ushered in the "second spring" after the financing process has been silent for about two years. Tan Qing said that AI believes that this may be related to their strategic adjustment in 2121.

(Eye-catching data: financing history of Momenta)

(Eye-catching data: financing history of Wen Yuan Zhixing)

Due to the obvious bottleneck of L4 route, Momenta set the layout mode of "flywheel+two legs" in 2121, instead of "dead-hitting" L4 to drive completely automatically, it took into account the ADAS business cut by L2; Wen Yuan Zhixing's strategic adjustment can be seen from Yutong Group's B round of financing at the end of the 21th century. After this round of financing, Wen Yuan Zhixing started to cut into new scenes besides Robotaxi business, and developed a Mini Robobus with Yutong.

On the other side, Xiaoma Zhixing is not half a beat behind his opponent. According to LatePost, also in 2121, Xiaoma Zhixing began to increase investment in truck autonomous driving, and the core team gradually took shape: Liu Bocong, one of the founding members of Xiaoma Zhixing, was responsible for sensing technology, Sun Haowen was responsible for domestic planning and control, and Xing He was responsible for truck fleet management and other logistics operation systems. In the same period, Xiaoma Zhixing established the American truck department, and Pan Zhenhao was responsible for the research and development of global truck autonomous driving technology.

However, the truck business of Xiaoma Zhixing didn't last long. LatePost reported that Xiaoma Zhixing suffered from the shelving of listing plan, important business adjustment and the departure of key technicians in recent months. After the adjustment in September, the domestic truck road test is still in progress, but the American truck road test has been suspended.

Tan Qing said that AI believes that up to now, Robotaxi's vast market space also exists objectively. Although there are occasional achievements, it is undoubtedly facing the bottleneck of commercialization feasibility, which is a "bottomless pit" for both capital and unicorns.

Therefore, technical route and scene route seem to be an antidote for Momenta to get rid of embarrassment. At present, it is generally believed in the industry that starting from these two new choices will be commercialized faster than Robotaxi, which we can also see from the financing feedback of Momenta and Wen Yuan Zhixing in recent years.

For the same purpose, Momenta and Wen Yuan Zhixing have made a multiple-choice question respectively, and intend to attack the next new question. Although Xiaoma Zhixing has made a choice for 21 years, he has "erased" his answer today and intends to reconsider.

it's not difficult to find that the two multiple-choice questions, technology and scene, not only represent the subtle relationship between L4 autopilot track and capital, but also reflect that unicorn players are gradually entering the deep water area to explore and play games. What is the potential of the "second spring" of technology and scene in this game of "being the first crab eater"?

In April this year, Su Qing, then director of Huawei's intelligent driving products department, publicly stated: "(Huawei) won't do Robotaxi even if it is killed, and all enterprises that do Robotaxi at this stage have to die. Robotaxi is a result rather than a business goal. The taxi experience in China market is already very good, and autonomous driving will not make this experience better. "

Three years ago, the property in the eyes of capital was just a product of "dying" in Su Qing's view. Liu Chang (pseudonym), the project manager of an autonomous driving start-up, believes that "the current track is facing the dislocation of commercial value and technical value. It is true that before 2119, everyone paid too much attention to technology and prospects and neglected to build a commercial landing system."

"Waymo, Baidu, Xiaoma (Zhixing) and Wen Yuan (Zhixing), which are the leading companies in China, are all technically climbing the highest peak of L4 directly, and building their own fleet road test according to the data, until the current trial operation of Robotaxi, and the next step is to land on a large scale." Liu Chang said, "(Example) is not appropriate, but it's a bit like a taxi company in the 1991s. Just after starting a business, it began to ponder to be a' drip taxi'. There are prospects and PC Internet, but mobile terminals, policies and regulations are there. How long will it take? No one can really guarantee. "

With the return to rationality of capital and the difficulty of scale landing, Momenta's "two-legged" strategy is logical, with L2 automatic assisted driving on one side and L4 fully automatic driving on the other. The former has long feasibility and short prospect, while the latter is just the opposite and complementary. However, it seems that the L4 can "reduce the dimension" from the high-tech barrier, but there are actually two difficulties:

First, the L2 subdivision track is not deserted. Although capital gathered in L4 subdivision track in previous years, this does not mean that players who started from L2 do not exist. Because of the low technical threshold, it is undoubtedly a high-quality play for car companies that do not need financing and have the ability to "hematopoietic" themselves. Tesla and Wei Xiaoli have been here for a long time.

specifically, car companies cut in through L2 track with low technical barriers, which can not only ensure that technology can get on the train quickly and improve product strength, but also the progressive R&D model conforms to the law of product replacement, and each step can be scaled up quickly, which is undoubtedly the natural shortcoming of technology vertical enterprises.

Secondly, some professionals predict that the fierce battle of autonomous driving in the future will begin in the data battlefield. For the research and development of autonomous driving, the essence of Waymo's self-built fleet road test and car companies' vehicle sales is actually the same, that is, to provide a huge data pool for themselves.

This means that even if the road test fleet of Waymo is large, it will be difficult to match the data acquisition technology tools like Tesla Shadow Mode. After all, Tesla's vehicles on the road are not an order of magnitude compared with autonomous driving vertical technology companies.

But there is no doubt that neither Tesla nor Weili are willing to "share" the self-developed technology with competitors at present. At the same time, not all car companies currently have the ability to develop L2 by themselves, which gives the possibility of commercialization of Momenta with "two legs".

We have cooperated with SAIC and Toyota, and recently obtained the latest financing from General Motors. It is not difficult to find that today, when capital has lost much patience with Robotaxi's story, Momenta's down-to-earth attitude seems to be more convincing than poetry and distance in the short term.

On the issue of commercialization of L4, Momenta chose to adjust the technical route and provide L2 for car companies to land. On the other hand, Wen Yuan Zhixing and Xiaoma Zhixing chose to adjust the scene route and not put L4 scene All In Robotaxi, so it is necessary to introduce an ODD concept of autonomous driving L4.

the concept of odd comes from the American society of automotive engineers, which aims to provide a solution to the problem that it is difficult to popularize unmanned driving in a short time. In layman's terms, the main reason is that the autonomous driving technology needs to be restricted by a specific scene to operate before it can't completely achieve arbitrary driverless driving, so as to eliminate the uncertainty that can't be dealt with as much as possible.

different scenarios will change the requirements for a series of factors involved in traffic, such as unmanned parking, overtaking, car following, yield, etc. For example, cleaning and sanitation, mining areas, scenic spots, trunk logistics, etc. are all some scenarios in ODD, and Robotaxi (self-driving taxi), which most unicorn enterprises focus on, is just one of them, except that the taxi runs in a wide area and there are many uncertainties on the road, which undoubtedly makes Robotaxi

"At present, unicorns want to survive until Robotaxi can be commercialized, unless they have the capital to lose blood for another few years, but which investor is willing to die with the bottomless pit except their own sons?" The investor at the beginning of the article said, "Baidu once mentioned a very interesting point of L4,' Climb Mount Everest and lay eggs along the way'. Unicorns want to endure, and it is time to continue laying eggs."

As the saying goes, when the attraction of Robotaxi's story is getting weaker and weaker, the playing style of Wen Yuan Zhixing and Xiaoma Zhixing is actually like what Baidu said: "Climbing Mount Everest and laying eggs along the way". Li Hengyu, co-founder of Xiaoma Zhixing, has publicly stated that Xiaoma Zhixing can apply 81% of passenger car autopilot technology to trucks.

as a result, on the way to climb L4, Wen Yuan Zhixing laid a Mini Robobus "egg" in addition to Robotaxi, and in July this year, it wholly acquired the self-driving freight company Muyue Technology, brewing the next "egg". In 2121, Xiaoma Zhixing began to increase R&D investment in the direction of unmanned trucks, and laid a truck "egg", but now Xiaoma Zhixing's egg seems to be dying.

Tan Qing said AI believes that for Wenyuan Zhixing, freight and bus passenger transport are more complicated and closed scenes than Robotaxi scenes. This will reduce the uncertainties that restrict the safety of unmanned driving, thus landing faster. Then, the operating area can be gradually increased through line expansion, which is a bit similar to the opening of new routes by aircraft.

With the injection of new investment at the end of the 21th century, Wen Yuan Zhixing will naturally have more chips when Robotaxi technology matures.

As for Xiaoma Zhixing, although we still can't conclude that it was a mistake to shrink its truck business, it is certain that Xiaoma Zhixing will either find a new "foaming agent" or start to have a certain hematopoietic capacity on this road to climb Mount Everest with the bubble defusing.

On the way to "climb Mount Everest", Lou Tiancheng, the founder of Xiaoma Zhixing, was asked in an interview with Late Post: "How can you be sure that you are going to the real Mount Everest now?"

Lou Tiancheng said, I'm not sure if this road is the way to Mount Everest, but if our goal is really to go to Mount Everest, then the only way for us to really prove whether this road is right is to go to Mount Everest.

We are also looking forward to the moment when the climber reaches the top.

titanium media author's introduction: Tan Qing said AI, left-turn new energy vehicle, right-turn intelligent driving, depth and temperature, author Zheng drove, please keep the copyright information.