In 2015, Google's AI (artificial intelligence) startup DeepMind released the artificial intelligence Go robot AlphaGo and fought against the then world Go champion Lee Sedol. In the end, the match ended with AlphaGo beating the pro nine-dan player 4-1.
After the sensational "man-machine war", the field of artificial intelligence has begun to receive widespread attention and has become a new track for the industry where talents and capital are rapidly flowing in.
According to the statistics of the "2020 Chinese Intelligent Industry Investment and Financing Report" of the Head Leopard Research Institute, as of February 2019, the number of AI-related enterprises in China was 745, accounting for about 21.7% of the world, and many of them were established after the "human-machine war" in 2015.
In recent years, with the rapid development of artificial intelligence, the industry has gradually differentiated, and artificial intelligence enterprises represented by SenseTime, Megvii, Yuncong, and Yitu have been born. These four companies have also been named the "FOUR LITTLE DRAGONS OF AI", and now they have all started their own listing steps.
On August 27 this year, SenseTime submitted a prospectus to the Hong Kong Stock Exchange. Since SenseTime's market share and financing scale rank first in the "AI Four Tigers", its listing in Hong Kong has attracted much market attention.
Red Star Capital Bureau will talk about whether this upcoming AI unicorn is a good company through SenseTime's prospectus, and whether it is a good business.
Infographic According to Visual China
(i)
3 and a half years loss of 24 billion
Where did SenseTime's money go?
SenseTime was founded by Dr. Xiaoou Tang. In 2014, Tang Xiaoou's team released research results, based on the original face recognition algorithm, with an accuracy rate of 98.52%, breaking through the human eye recognition capability (97.53%) for the first time in the world.
Tang Xiaoou became so famous that soon IDG Capital partner Niu Kuiguang came to visit and invest tens of millions of dollars. In October 2014, SenseTime was officially established.
For the future of SenseTime, whether it is product landing or liquidity, Tang Xiaoou is full of confidence. At the 2018 SenseTime Artificial Intelligence Summit, Tang Xiaoou said: "Academic things, if they can't land, there is no use, fortunately, we are not a company that burns money, it is a company that can make money, it can be self-financing, our financing is not used to burn, but to do great things." ”
Xiaoou Tang at the 2018 SenseTime Artificial Intelligence Summit
With the hard power of the academic school, SenseTime has always been regarded by capital as "fragrant food".
Public information shows that from 2015 to 2021, SenseTime has raised 12 rounds of financing, totaling $5.2 billion.
According to the prospectus, as of June 30, SenseTime had RMB8.926 billion in cash and cash equivalents. As the "darling" of capital, SenseTime does not seem to be short of money at present.
Although it is not very bad money, through the prospectus, it is not difficult to see that SenseTime, as an AI giant, is a veritable "gold-devouring beast".
According to the prospectus, from 2018 to the first half of 2021, SenseTime's net losses were 3.438 billion yuan (RMB), 4.968 billion yuan, 12.158 billion yuan and 3.713 billion yuan, with a cumulative loss of 24.277 billion yuan. After deducting non-recurring gains and losses, from 2018 to the first half of 2021, SenseTime's adjusted net losses were RMB221 million, RMB1.037 billion, RMB878 million and RMB726 million, respectively, with a cumulative non-net loss of RMB2.86 billion.
On the one hand, there is a super financing "supply", on the other hand, it is a "bill" of annual losses, where is senseTime's money spent?
The answer is not difficult to find, as a technology-based enterprise, "people" are often the key to determining the rise and fall.
According to the prospectus, from 2018 to 2020, SenseTime's R&D investment was 849 million yuan, 1.916 billion yuan and 2.454 billion yuan, accounting for 45.9%, 63.3% and 71.3% of the current revenue, respectively; in the first half of 2021, it reached 1.772 billion yuan, accounting for 107.3% of the revenue in the first half of the year. At present, senseTime's R&D investment has reached 6.991 billion yuan.
Source: Company earnings
SenseTime's burning money is first attributed to huge R&D expenditures, of which the salary expenses of R&D personnel account for the majority of R&D expenditures.
In the three and a half years from 2018 to the first half of 2021, senseTime's expenditure on salaries and benefits (including equity incentives) for R&D personnel increased year by year. In 2018, the salary and welfare of R&D personnel was 513 million yuan, accounting for 60.4% of R&D expenditure; in 2020, the salary and welfare of R&D personnel was 1.569 billion yuan, accounting for 63.9% of R&D expenditure; and the salary paid in the first half of 2021 was 1.285 billion yuan, accounting for 72.5% of R&D expenditure.
Source: Prospectus
Although most of them are issued in the form of shares, and the recent rapid growth in the fair value of the company's shares has increased the salary level of employees on the books, it is difficult to hide the fact that SenseTime has made a big deal with R&D personnel.
In addition to R&D expenditure, SenseTime's administrative expenses are also very "huge".
In 2020, sensetime's administrative employee salaries totaled $931 million, accounting for 58.5% of the total annual administrative expenses; by the first half of 2021, administrative employee salaries totaled $1.148 billion, accounting for 79.7% of total administrative expenses.
Most of them are still issued in the form of shares, and in the first half of 2021, SenseTime once again paid up to 1.188 billion yuan in compensation to directors on a share basis, and the vast majority of which was included in administrative expenses, which also pushed up the salary expenses of administrative employees in the statement.
Judging from the prospectus, the huge R&D and administrative expenses are the main reasons for SenseTime's "burning money", and these two expenses are directly related to "people".
(ii)
Split the revenue structure
Is SenseTime Solid?
According to the prospectus, SenseTime's business is mainly divided into four parts: smart business, smart city, smart life, and smart cars.
Among them, smart business and smart city are currently the main sources of revenue of SenseTime, accounting for about 80%, but the growth rate of these two major revenue sectors has slowed down.
According to the prospectus, SenseTime's smart business revenue from 2018 to 2020 was 8.53 billion yuan, 12.03 billion yuan and 14.85 billion yuan, respectively, and the growth rate in 2019 and 2020 was 41% and 23%, respectively; the revenue of smart cities from 2018 to 2020 was 5.3 billion yuan, 12.7 billion yuan and 13.69 billion yuan, respectively, and the growth rate in 2019 and 2020 was 139% and 7%, respectively.
In terms of segmentation, the two main revenue businesses of SenseTime also face certain limitations.
The first is the smart business business, which has a significantly lower number of customers than expected. The number of customers in 2018 was 539, and by 2019, it increased to 834, a growth rate of 54%, while the number of customers in 2020 was only 848, a growth rate of only 1%.
The second is the smart city business, although the total number of customers has risen, but the smart city mainly provides government departments with the detection of public facilities and tracking the impact of natural disasters, which also means that the business cycle is long, and the payment cycle is generally relatively long, so it may lead to a certain impact on the cash flow of enterprises.
Perhaps considering that the two main businesses will face greater challenges in the future, SenseTime has also begun to make continuous efforts in smart life and smart cars to find breakthroughs.
According to the prospectus, in the smart life business, SenseTime has built a multi-layered infrastructure to empower IoT devices and drive metaverse to enhance the end-user experience. As of June 30, 2021, SenseME and SenseMARS have enabled more than 450 million mobile phones and more than 200 mobile applications.
In the smart car business, SenseTime said that as of June 30, 2021, it has cooperated with more than 30 car companies and has been selected as a supplier of more than 50 models, and will supply more than 20 million cars with shadow products in the next few years.
However, for the exploration of these two businesses, rational investors may not give sensetime positive feedback. After all, the concept of metacosmology is still a long way from being illuminated into reality, and the commercialization of smart cars is not achieved overnight.
(iii)
Coming soon
How much imagination does SenseTime have?
SenseTime has played its best performance at the best time, has been attracting much attention, and has also won the "favor" of capital. However, the long-term loss state and the continuous "bottomless" research and development investment may also make the market have some doubts about the upcoming AI unicorn.
In fact, not only SenseTime, the "AI Four Little Dragons" are currently in a state of loss, and they are eager to go public.
The problems facing SenseTime today, like the rest of the "three brothers", are difficult to escape external troubles and internal worries.
From the perspective of external troubles, with the continuous development of technology, the AI industry has lost its former mystery, market competition has begun to intensify, and the strength of opponents should not be underestimated. Large ICT integrated manufacturers represented by Huawei, as well as Internet giants such as Ali, Tencent, Baidu, and hardware leading enterprises such as Hikvision and Dahua Shares, have entered the AI track, and SenseTime may not occupy an absolute advantage.
From the perspective of internal worries, the banner of science and technology is held high, but how to land in the end may be the core internal contradiction facing SenseTime today. SenseTime started in the field of computer vision, and is still relatively limited in application scenarios, mainly based on security and finance. For SenseTime, the future "tune" is too high, easy to deviate from reality, but the "tune" is too low, and it will make itself not have a competitive advantage.
SenseTime has successfully caught up with the outlet of artificial intelligence, but the real spring of artificial intelligence has not yet arrived. Today's Shang Tang is nervous and urgent, because the real big test is approaching step by step.
Red Star News reporter Yu Yao Liu Mi
Responsible editor Ren Zhijiang Editor Yu Dongmei
(Download Red Star News, there are prizes for the newspaper!) ）