Image source @ Visual China
Wen | talk about AI, and the author | Zheng driving
Recently, SenseTime released its first financial report on the listing.
According to the financial report information, SenseTime's total revenue in 2021 was 4.7 billion yuan, an increase of 36.4% over 3.446 billion yuan in 2020. However, the loss situation is still serious, with an annual loss of 17.177 billion yuan, an increase of 41.3%. Excluding employee compensation incentives, listing expenses, financial fair value, etc., the adjusted net loss was 1.418 billion yuan, an increase of 61.5% year-on-year.
As one of the few AI companies on the Hong Kong Stock Exchange, SenseTime's earnings report still seems to release a tone that is developing first at this stage.
A noteworthy move is that on the same day that SenseTime's first financial report was released, Sun Dapeng, vice president of SenseTime, attended the 2022 China Electric Vehicle 100 Forum and expressed SenseTime's thinking on the smart car business and the planning of automotive products.
Sun Dapeng pointed out that the current business of SenseTime can basically be summarized as "one platform, four pillars". "One platform" is the artificial intelligence infrastructure, equipped with a self-developed training framework. The "four pillars" are the four business sectors formed on this basis, namely smart business, smart city, smart life and smart car.
Judging from the financial report information, although SenseTime's smart car business accounted for the lowest proportion of revenue last year, it contracted slightly compared with 2020. However, SenseTime said that in the first quarter of this year, SenseTime has completed the upgrading of its organizational structure and established a smart car business group, which will make greater efforts to develop the smart car business.
On the other hand, combined with the current trend of electrification and intelligent transformation of the automotive industry, the automotive business also has a huge potential for untapped value to a certain extent for SenseTime.
So in the face of the still expanding loss situation, can smart cars become the next fortress carrying SenseTime's future imagination?
The financial report is about the star sea of cars
In the 51-page financial report of SenseTime, although the relevant data of smart cars is not eye-catching, the word automobile is mentioned 18 times, compared with the business mentioned 14 times and the life mentioned 11 times, which are only a lot more.
The car track seems to have vaguely let us see the sea of stars of SenseTime about the future. Why? Tan Qing said that AI believes that there are mainly the following two points:
Ai fresh air from technology to mass production
In the past two years, the commercialization of the application layer has gradually become one of the hottest topics in the field of AI, and the development focus of many domestic AI companies has naturally focused on the terminal application layer.
Such a trend of change is not difficult to understand, take automatic driving, when the track has just been pushed to the outlet of the technology companies are often the most popular, but after a bottleneck, technical barriers are likely to collapse rapidly, and even by many latecomers to catch up.
Compared with the hot money surge in previous years, the financing environment of the L4 high-end autonomous driving track has begun to be cold, and various players have sought new barriers on the landing, which has already allowed us to see such a trend.
Looking further at the financial report, in 2021, SenseTime's total revenue achieved an increase of 36.4% compared with 2020, and SenseTime believes that the accelerated integration of AI technology and the real economy has brought accelerated growth in its annual performance.
This sentence seems to point directly to the arrival of AI mass production opportunities, which is bringing new vitality to SenseTime.
So now the pressure has come to the head of SenseTime, The smart business with the largest revenue in 2020 is not as good as the smart city in 2021, the proportion has fallen from 43.1% to 41.7%, and the smart life is even more in terms of income and proportion than in 2020, creating RMB revenue of 420 million yuan in 2021, accounting for 8.8%.
From an industry perspective, Google, Facebook, Microsoft, BAT, etc. among the tech giants seem to be building their own scalable and iterable AI open source frameworks.
To some extent, if technology open source becomes a tacit understanding of technology giants, it may bring a blow to enterprises in the vertical technology layer field, and "flood" the technology "flood" that is in the hands of a small number of people.
This means that pure technology is actually difficult to be the real barrier of AI companies, and only by continuously deepening the deep integration of technology and industry is king, so SenseTime needs to take root as soon as possible in the scene that has been cut.
However, at present, SenseTime's customer base attributes in more scenarios are still 2B or 2G, just like products with strong correlation with computer vision, in our cognition, the unit price of customers may be relatively high and low frequency, although the scene application space is wide, but the overall landing scene is limited. SenseTime serves more businesses focusing on the smart life of 2C enterprises, but last year's revenue accounted for less than 10%.
This seems to let us see that if SenseTime wants to continue to maintain sustainable growth momentum in the future, it also needs to constantly look for new scenarios. Then the automobile industry, which is in the midst of a wave of intelligent revolution, is obviously a rather high-quality target.
More importantly, the car is not a simple superimposed scene increment, we may wish to further look at it from a strategic level.
Business tandem value in automotive scenarios
According to the financial report, SenseTime's R&D expenditure in 2021 increased by 47% from RMB2.454 billion in 2020 to RMB3.614 billion in 2021.
SenseTime said the further increase in R&D spending was mainly due to an increase in employee benefits expenses, on the one hand, the expansion of the R&D team, and on the other hand, an increase in the share-based compensation expenses of R&D personnel.
However, the current business layout of SenseTime does look a bit large and scattered at first glance, and it seems that it is difficult not to focus on the suspicion of loss. After all, from the business of ToB to the life of ToC to the city of ToG, SenseTime can almost be said to have laid an all-encompassing AI network.
At this time, the input-output ratio seems to be a problem that Needs to be faced more than others for SenseTime. Some people may question whether the booth is already large and the investment in research and development is increasing day by day, and then it will be necessary to enter the intelligent car business in a big way, will it bury more hidden dangers for the financial reports in the next few years?
Tan Qing said that AI believes that in fact, it is just the opposite, from the perspective of business synergy, smart cars are not actually SenseTime's new guns, but the continuous fuel at the strategic level. The automobile business has the potential to feed back its overall business territory, which can drive its three existing cruise ships of business, life and city to sail more smoothly.
Specifically, this is actually what Sun Dapeng mentioned, intelligent cars "combined with smart business to form automatic connection products, which can effectively reach commercial hubs and promote business model transformation; cross-form smart transportation products with smart cities, so that the vehicle-road collaboration system can effectively integrate various business scenarios and applications of smart transportation, bringing more choices for travel; combined with smart life, the entertainment application of the mobile phone can be effectively applied to the entire cabin, so that the passenger experience is more beautiful. ”
It is not difficult to find that entering the automotive intelligence can indeed form a synergy with other businesses of SenseTime, but along this line of thinking, you need to pay attention to a point, if you want to maximize the release of this synergy, it is likely to mean that in the layout of the field of automotive intelligence, SenseTime still needs to carry out a large and scattered global attack.
Car intelligent fish and bear paws, Synthmus all want?
What are SenseTime's layouts in the field of smart cars?
Sun Dapeng said that SenseTime has formed three product sequences on the engine that empowers the ultimate shadow, which bets on the intelligence of automobiles:
The first product sequence is the intelligent cabin, which is more to provide solutions for intelligent and interactive entertainment in the car; the second product sequence is mass-produced intelligent driving, mainly for front-loading mass-produced customers; the third product sequence is the intelligent driving connection series, mainly L4 level automatic driving.
In terms of automotive intelligence elements, SenseTime from intelligent cockpit to ADAS assisted driving to L4 high-end automatic driving can be said to be involved in almost all important links in the second half of the entire intelligent electric vehicle. From the perspective of technical path, SenseTime actually does not only bet on bicycle intelligence, but also pays a certain degree of attention to vehicle-road collaboration.
Unsurprisingly, SenseTime's layout in automotive intelligence can be described as similar to its macro business layout, that is, it is more and more scattered, and it seems to want to cultivate the smart car business into an all-round player.
From the perspective of the macro autonomous driving track, the same listed unicorn Tucson Future, or the head player such as Xiaoma Zhixing, in fact, rarely has such an all-encompassing layout model.
But it is true that in the spirit of big talents do not choose, I want all the original intention, let's first look at the layout method of SenseTime, what are the advantages.
First of all, it is to be able to implement the ecology of the business layout, which will not be repeated too much. As we mentioned earlier, if SenseTime wants to maximize the value of the smart car business, it is difficult to position it as an absolutely isolated existence.
On the other hand, AI technology actually has a lot of underlying reuse value, and the global layout can improve the technology reuse rate, and SenseTime seems to know this well. For example, Tesla still has challenges in building cars, but it still has to start to have an in-depth layout on robots.
However, some of the giants are large and scattered on Shang Tang, and the hidden worries of the two aspects cannot be ignored:
The real body of the unicorn under the giant mold
First of all, an obvious hidden worry is that the continuous loss of money in the short term seems inevitable for SenseTime, after all, in the field of smart cars, there are very few players who can achieve large-scale profitability by relying on this business alone.
Not to mention that vehicle-road collaboration, which is still in the early stage of development of the program path, before mass production, there is no doubt that it will go through a complex process of industrial collaboration and mutual running-in.
Let's look at players like Baidu or Huawei, who are not at all soft in betting on car intelligence, the former from smart city to vehicle-road collaboration, from L4's high-level self-research to personally building cars, almost everything; the latter is a global range of software and hardware, intelligent cockpit, ADAS, and even lidar and other perception hardware is also doing.
Compared with such giant players, the shortcomings of SenseTime are obvious, that is, the giants have a blood-making business that can be inherited from the cash cow business, and this weapon of SenseTime is still relatively missing in the short term.
Therefore, Tan Qing said that AI believes that if the intelligent fish and bear paws of automobiles want to have both, SenseTime may continue to dormant, then the continuous losses that follow may be the price to be paid for its reciprocity.
Delicate market environment
Looking in a clear direction, the ADAS market and the mass production capacity of the smart cockpit have become more and more impressive, not as elusive as the L4 or the current stage of vehicle-road collaboration, can SenseTime first stand on the first leg here?
Admittedly, market opportunities have emerged, but SenseTime has not embarked on a smooth road.
Tan Qing said that AI believes that the current ADAS and smart cockpit market is a very clear problem is that the competition is quite fierce. If SenseTime wants to stick a kick in, it needs to first think about the delicate relationship between the three industries, or the camp of enemies or friends, namely OEMs, technology giants, and startups.
Since last year, more and more OEMs have begun to relax on the fat meat of assisted driving, such as SAIC, which speaks of the theory of body and soul, or BYD, which has always insisted on self-research, to a certain extent, and has begun to find "foreign aid".
Tianyancha APP information shows that in momenta's C round and C+ round of financing, SAIC motor has appeared, and BYD has announced cooperation with Baidu and NVIDIA this year.
But that doesn't mean that unicorns like SenseTime can easily get a piece of the market that's waning.
First of all, start-ups are often the target of more high-quality cooperation between OEMs at this stage. Although there is a trend from self-research to cooperation, the current cooperation of main engine factories such as GAC and SAIC is more or less vigilant, and most of them cooperate by participating in technology vertical start-ups, so as to grasp the initiative as much as possible.
Secondly, packaged technology solutions are also the current favored cooperation methods of many OEMs, and the reason why Huawei has been able to cooperate with new brands such as Cyris, Avita, AITO and so on in recent years is partly because it has the "one-stop" service capability from intelligent cockpit to auxiliary driving, from hardware to software.
But more importantly, Huawei's deep packaging cooperation, the "one-stop" concept does not stop at products, in addition to providing partners with brand endorsements with great marketing value and a huge offline sales grid, which is undoubtedly missing from SenseTime.
It is not difficult to find that in these two mainstream cooperative relations, a listed company of such a large size as SenseTime will actually be slightly embarrassed to a certain extent, and therefore, the independent car companies that are now relatively gradually open may be the opportunities for SenseTime's mass production intelligence.
After all, BYD's cooperation with Baidu, or The cooperation between Extreme Kr and Mobileye, has already released this signal to us to some extent.
So can SenseTime seize the opportunities of smart cars in the future?
According to the financial report, SenseTime has signed more than 30 customer contracts with cooperative car companies in 2021, and the number of pre-installation fixed points for cooperation has reached 23 million units. In contrast, in this year's financial report, the revenue of smart cars still accounts for less than 4% of the overall business.
Perhaps today, it is still difficult to give a positive prediction.
However, in the dangerous market environment, the most brilliant life can often be bred, and the only thing we can do is to bless Theorem, which is rooted in technology and grows freely, and continue to fight in the sea of stars of smart cars, waiting for the flowers to blossom.