Layout of automatic driving, Cheng Wei and Didi "dilemma"

Text: Tan Qing said AI Author: Zheng Kai

In the wail of L4, Didi's automatic driving is belated.

On April 13, Didi announced a series of new products and services, including the first future service concept car DiDi Neuron and the first autonomous driving automatic operation and maintenance center.

Perhaps in order to avoid too much discussion from the media, Cheng Wei did not speak on stage at the press conference, but just sat in the corner and quietly watched the press conference from beginning to end. Without the founder's speech that rebuked Fang Xuan, the press conference itself seemed to have less flavor.

They all came, and they didn't go on stage to say two words, which seemed to reveal some dilemmas.

Layout Robotaxi, Didi dilemma

The "dilemma" is not only Cheng Wei, but also Didi autonomous driving.

From the perspective of the conference itself, the core is still the story of L4 automatic driving, whether it is a concept car or an automatic operation and maintenance center, or the "old story".

Cheng Wei once told the media that there are only one or two places in the autonomous driving market, and there is no third place, just like Windows and Communix, Android and iOS, outside Google, and Didi hopes to be another player who can survive.

Whether Didi can be the top two in the industry is still unknown, but L4 seems to be a "must-have" for Didi to do.

However, the embarrassing thing is that the story of L4 in the past two years is no longer fragrant.

First, Mobileye's valuation fell by two-thirds, Waymo's valuation fell from a maximum of $175 billion to $30 billion, the venture capital of the self-driving industry shrank seriously, and unprofitable driverless projects all encountered survival problems.

Since then, Waymo has begun to lay off employees, rumors that Pony.ai has also laid off employees has been officially denied, and Tesla's self-driving team has also laid off hundreds of employees...

After the industry's winter layoffs, Didi's automatic driving is in a dilemma.

Tan Qing said that AI believes that for Didi, doing high-level autonomous driving is not a Normandy landing-style attack, but a Stalingrad-style defense.

The landing of high-end autonomous driving is bound to trigger a new round of structural adjustment in the ride-hailing business.

At present, there are three structural adjustments in the ride-hailing business:

The first is the adjustment of the matching method: from the "car looking for people" of traditional taxis to the "people looking for cars" under the matching of supply and demand.

The second is supply-side adjustment: the platform's offline fulfillment capacity and market share competition.

The third is cost restructuring: cost structural changes triggered by autonomous driving technology.

The adjustment of the first two stages has been basically completed, and after the second adjustment, players such as Meituan, Didi, and Xiandao have formed their own barriers and boundaries.

The essence of the platform economy is network effect, it is difficult for players in the same track to subvert other players, and the industry is actually fighting for cost advantages in the end.

Among the cost of online ride-hailing, the vehicle is a fixed cost, the driver's income and operating expenses are elastic costs, but the effect of operating cost reduction is limited, and the most expensive cost of online ride-hailing is actually the labor cost.

Taking Meituan's data in 2018 as an example, its ride-hailing driver cost was 4.463 billion. If this part of the cost is converted into revenue, then the increase in profit is quite good. Therefore, after the popularization of autonomous driving and the removal of driver costs, the overall cost structure of ride-hailing will be completely changed, which brings huge imagination.

In the face of such an increment, if nothing is done, then Didi's basic disc may be shaken.

Therefore, high-level automatic driving, Didi must do it.

As early as 2016, Didi established an autonomous driving research and development department. In August 2019, Didi announced the spin-off of its autonomous driving division and the establishment of an independent company, with Zhang Bo, then Didi's CTO as CEO.

In the same year, Didi conducted a small-scale autonomous driving test in Shanghai's Jiading District, and proposed the goal of becoming "China's first commercial autonomous driving ride-hailing service + hybrid dispatch service" platform.

The products at this conference are the result of Didi's hard work in recent years.

But the question is, if the direction is wrong, no matter how hard you work, will the capital market pay for it?

Didi's automatic driving is equivalent to changing the track, and in the face of the absolute cost advantage of automatic driving, Didi's basic disk needs to be restructured. The foundation of the re-architecture lies in technology research and development and route selection, which is not only the competition of business models, but also the competition of technological innovation.

Betting on L4 on the technical route and Robotax on the commercial route seems to be two good moves, but it is actually two "smelly chess". The strategic route is seriously disconnected from the development of the industry.

The reason behind this may be that management is disconnected from the commercialization of autonomous driving, or it may be that it relies too heavily on the established technical roadmap. But whatever the reason, it may not be a good thing.

Judging from the press conference, whether it is the autonomous driving automatic operation and maintenance center or the first future service concept car DiDi Neuron, how to look like going to L4.

It is undeniable that if these solutions can really be scaled and commercialized, the imagination space is indeed large.

But the problem is that whether it is Baidu Radish Run, which Didi wants to target, or other players who do Robotaxi, they are no longer "Britney" in the eyes of capital.

Waymo's valuation has plummeted, and industry players have laid off workers to survive, showing that the capital market not only wants a "good story", but also a "story" that can generate short-term gains, at least until L4 and L5 land.

What is the situation of the L4 route, I believe that friends who are familiar with the autonomous driving industry are very well aware.

The large-scale and commercialization of Robotaxi is even more remote.

Today's regional commercialization, not so much landing, but a commercial demo, or to show the market, how many Robotaxi projects can really earn money from C-end users? Robotaxi to earn money, in fact, is to earn the cost of the salary of the ride-hailing driver, now each Robotaxi has at least one security officer, so where can the cost be low?

In addition, from the perspective of commercial operations, regional operations mean that only the needs of short-haul users can be met, and the significance of pilots is self-evident in the point-to-point operation of some robotaxi vehicles.

The reality is that the regionally operated Robotaxi is still a "model room". The industry is far from the real landing of L4, and the landing of Robotaxi is even farther, as short as 10 years and as long as 20 years.

The capital market has shown the attitude of L4 + Robotaxi, at this time, to bet on Robotaxi, can it regain favor? I'm afraid it's also difficult.

Next, how to "play flowers" in the commercialization of autonomous driving may be what the capital market wants to see.

In a dilemma, independent IPO or "new life"?

Judging from Didi's current situation, it seems more important to regain confidence in the capital market after the turmoil.

Tianyan inspection shows that Didi's two most recent financings are in 2021, and whether it can be truly commercialized next is crucial to win the trust of capital.

After Didi applied for delisting, there have been rumors that Didi wants to go public in Hong Kong.

Going to the secondary market again, whether investors can continue to pay is a big question.

An awkward situation lies in front of Didi and Cheng Wei: the story of L4 and Robotaxi is no longer sexy, and the market does not buy it, but not betting on autonomous driving may mean that it will lag behind in technology, and lagging behind will be "beaten".

Continue to Robotaxi to the black, is likely to be dragged into another cost trap, although the ride-hailing business is profitable, but the risk is still there, whether it can support long-term research and development investment, investors actually have no bottom.

From a strategic point of view, Didi's automatic driving seems to be somewhat "kidnapped" by the ride-hailing business.

For Xiaoma, Wen Yuan, driverless is entrepreneurship, and in the end, even if it can't be done, it is nothing more than the ideal being shattered, re-positioning, and changing the track, such as becoming a T1 supplier of the main engine factory, which is also a good choice.

But for Didi, automatic driving is related to the basic disk, whether L4 is reliable or not, it must be hardened, strategically cannot retreat, and the technology landing is a bit "chicken ribs", it is difficult to do a compromise. That is to say, in terms of strategy and tactics, Didi is actually in a "passive".

Other enterprises, grabbing the L4 track, layout Robotaxi, Robobus and other tracks are "fighting rivers and mountains", and Didi is "guarding rivers and mountains". It's hard to cook, it's easy to know at a glance.

In fact, judging from the press conference, there are also some clues that reflect Didi's "passivity".

From the perspective of time nodes, if last year took advantage of the surge in AI concepts, the concept of autonomous driving rose to this conference, the effect may be better, and now it may still be a little late.

Today, Didi, which has not spoken for two years, is old, and whether it can keep up with the "variations" of the industry is a test of management's ability to respond.

Tan Qing said that AI believes that passive variations are not as good as active composition. The independent IPO of the autonomous driving business may be a "new life" opportunity.

First, the listing boom in the autonomous driving industry is still ongoing.

In February, Hesai Technology went public, and on April 4, Zhixing Automotive Technology submitted a prospectus for Hong Kong stocks, in addition to news that Wenyuan, Moment and other companies are seeking IPOs. In the commercial vehicle sector, start-up Veni also received financing.

In the AI cycle, autonomous driving is ushering in a new cycle.

Second, TO C businesses such as Robotaxi are not profitable, so commercialization depends on TOB.

The biggest advantage of Didi autonomous driving is its operation and performance capabilities.

To put it bluntly, there are customer needs and landing scenarios, but the difficulty is technology, now there is no L4, regulations and so on have not been popularized, when the L5 technology that can be implemented on a large scale will be made, no one knows.

To live in the present and gain market recognition, it is necessary to rely on To B to commercialize.

At the press conference, Didi came up with hardware solutions: "Beiyao Beta" lidar and three-domain fusion computing platform "Orca killer whale". These may be what the market is really focused on.

However, there are many challenges.

For example, as a "layman" in the eyes of the automotive industry, how to break into the supply ecology of car companies? In the face of car companies that require engineering capabilities and mass production capabilities, how to meet the needs of mass production?

Mass production capacity may still be the weakest link in Didi's business chain, and it is difficult to solve the problem of mass production due to the lack of sufficient industry resources and scale effects.

Previous cooperation with BYD and GAC Aion may bring some useful experience, but whether it can do a good job in mass production is still crucial.

In fact, the income generated by other players in the industry mainly relies on serving car companies, doing L2 on-board projects, and road coordination and other To G landing projects. In addition, autonomous driving is also constantly expanding other landing scenarios such as terminals, ports, trunk line logistics, etc. These areas are easier to land and easier to obtain commercial benefits.

In this regard, Didi also explores.

At the press conference, Didi announced that it has expanded its Level 4 autonomous passenger car solution to the field of trunk logistics and freight, and launched Didi's autonomous driving freight KargoBot.

According to Wei Junqing, head of Didi's autonomous driving innovation business, KargoBot currently has more than 100 self-driving trucks and has started normal trial operation between Tianjin and Inner Mongolia. As of March this year, KargoBot's cumulative logistics revenue has exceeded 100 million yuan.

The data shows that Didi autonomous driving has taken steps in the commercial exploration of scenarios, and from the perspective of commercialization, there are also some achievements.

This is quite important.

For Didi, the autonomous driving business has a "back road" L4 driverless landing, and there is also room for commercialization of autonomous driving. At the very least, the future will not be left behind.

In addition, this also means a pragmatic attitude, to do automatic driving not only to have "poetry and distance", but also to have commercial "rice, oil and salt".

The "plank road" of "Ming Xiu" unmanned driving, and the "Chen Cang" of "dark crossing" freight logistics may be the successful method for the commercialization of Didi's autonomous driving.

Written at the end:

This Didi autonomous driving open day is also quite meaningful to the industry.

After a long silence, this is the first official and large-scale "announced return" of Didi, and after Cheng Wei sat down, the outside world also expected Didi after the storm to return to everyone's attention again.

As a giant in the travel industry, Didi's progress in autonomous driving is also a rare comfort.

At a time when L4 is in a depression and Robotaxi has no capital appeal, unicorns such as Wen Yuan and Xiaoma need industry giants to turn the tide. And Didi's insistence on Robotaxi, what kind of fruit will be, it still takes time to give an answer.

Tan Qing said that AI production

Unauthorized reproduction is not permitted