Text | Good looking business, the author | Monday, the editor | peace of mind
From 2020 to the first half of 2021, SenseTime's revenue was RMB3.446 billion and RMB1.652 billion, up 13.8% and 91.8% year-on-year, respectively. According to media reports, the current valuation of SenseTime is 12 billion US dollars, and the IPO has raised at least 2 billion US dollars, and the total valuation will exceed 100 billion yuan.
The biggest IPO in the field of artificial intelligence is coming.
On August 27, SenseTime submitted a prospectus to the Hong Kong Stock Exchange. At this point, the "AI Four Little Dragons" have finally gathered on the road to IPO.
SenseTime, Megvii, Yitu Andun and Yuncong are known as the "FOUR LITTLE DRAGONs of AI" by the market.
Previously, megvii, YITU and Yuncong's listing on the Sci-Tech Innovation Board has made different progress: Megvii recently replied to the second round of inquiries of the Shanghai Stock Exchange; Yitu's SCI-Tech Board IPO folded and was exposed to consider Hong Kong IPOs; and Yun's initial listing from the Sci-Tech Innovation Board has been approved.
Although SenseTime was the latest to go public, it is the absolute leader of the four companies in terms of revenue scale, research and development strength, and valuation.
From 2018 to 2020, SenseTime achieved a total revenue of 8.326 billion yuan, which is 2.38 times that of Megvii and more than 4 times that of Yuncong; Yitu only disclosed the revenue as of the first half of 2020, and roughly estimated that its revenue in the past three years was less than 1/4 of SenseTime.
SenseTime also has an absolutely leading R&D strength. As of June, SenseTime had a R&D team of 3,593 people, and they had an intellectual property portfolio of 8,123, including 4,169 patent assets in China and 3,954 overseas.
Sullivan's report argues that as of June 30 this year, SenseTime has one of the largest portfolios of invention patents in the AI industry in Asia.
Of course, this also comes at a price, and SenseTime has invested huge research and development costs over the years. From 2018 to 2020, its R&D investment exceeded 5.2 billion yuan; in the first half of this year alone, it reached 1.77 billion yuan, which is the largest R&D investment among the four small tigers.
The investment is increasing sharply, and the loss is getting bigger and bigger. From 2018 to the first half of this year, SenseTime lost 24.272 billion yuan in three and a half years, ranking first in the scale of losses. In the three years from 2018 to 2020, Megvii accumulated losses of 12.77 billion yuan, and Cloud lost more than 2.6 billion yuan, according to the chart to the first two and a half years of 2020 loss of 6.1 billion yuan. However, these losses include fair value losses on preferred stock.
SenseTime has to support huge investments, and one of the important sources of funding is continuous financing. Since its inception in 2014, SenseTime has completed a total of 12 rounds of financing, totaling US$5.2 billion (about 33.658 billion yuan), with the largest number and amount of financing among the "AI Tigers".
According to media reports, the current valuation of SenseTime is 12 billion US dollars, and the IPO has raised at least 2 billion US dollars, and the total valuation will exceed 100 billion yuan.
In 2018-2020, the total revenue of less than 10 billion yuan in the three years is less than 10 billion yuan, what does SenseForce rely on to support the valuation of 100 billion yuan?
<h2>An IPO that is imperative</h2>
What does filing an IPO application now mean for SenseTime?
First, let's take a look at SenseTime's current operation and capital situation.
After adjustment, from 2018 to 2020, SenseTime's net loss was 221 million yuan, 1.037 billion yuan and 878 million yuan, respectively. In the first half of this year, SenseTime lost 726 million yuan, almost catching up with the level of the whole of last year. In 2021, SenseTime is unlikely to achieve profitability.
Among the risk factors in the prospectus, SenseTime mentioned that "we may not be able to achieve or maintain profitability in the future".
Net debt also exposes SenseTime to liquidity risk. As of June 30 this year, SenseTime's net debt was RMB22.961 billion.
As of June 30, 2021, SenseTime's cash and cash equivalents were only $8,926 million. In the next step, SenseTime still needs a lot of external financing to maintain its competitive advantage and rapid development.
Over the past few years, SenseTime has relied primarily on equity financing, bond financing and operating activities to meet cash flow needs. SenseTime, which has not yet achieved profitability, still needs equity financing to obtain development funds.
After all, it takes a lot of money to develop it alone. From the past few years, senseTime's R&D expenses have increased year by year, from 45.9% in 2018 to 63.3% in 2019 and to 71.3% in 2020; by the first half of this year, it has reached 107.3%.
SenseTime's large investment in research and development still can't stop. In this IPO, SenseTime plans to use about 60.0% of the funds raised to enhance R&D capabilities.
In the primary market, it is already very difficult for SenseTime to carry out equity financing. Previously, SenseTime has completed 12 rounds of financing in the primary market, totaling $5.2 billion.
SenseTime's financing history, image source: prospectus
As early as 2020, SenseTime continued to report that it was seeking a new round of financing, but it did not make substantial progress for a long time.
Judging from the financing process disclosed in the prospectus, SenseTime basically maintained the rhythm of financing rounds every few months from 2016 to the first half of 2019. However, from March 2019 to July 2020, It took SenseTime nearly a year and a half to finalize the subscription of the D series financing, and the pace of financing was significantly slower than in previous years.
According to the feedback of PE investors, the enthusiasm for AI investment in the primary market is fading, and AI unicorns have experienced multiple rounds of financing in the early stage, and the demand for investors to exit is becoming increasingly strong.
The primary market's attitude towards AI unicorns has gradually become rational from the initial enthusiasm, and now they are more interested in the results of commercialization.
Striving to promote revenue growth is the common state of the AI Four Dragons in recent years.
According to LatePost, in 2020, SenseTime CEO Xu Li set two MAJOR OKR goals for the company, one of which is sustained performance growth.
Xu Bing, co-founder of SenseTime, also mentioned, "We have paid special attention to the ability of business from the first day of establishment, any technology company, no matter what you do, the ultimate embodiment is still a business ability, low dimensional, is the ability to manage cash flow, you can not balance income and cost well." ”
However, judging from the financial performance of the AI Four Dragons, commercialization is still not easy.
They generally present the characteristics of slow technology monetization, low gross profit margin, and slow customer payment collection, which are also constantly diluting the aura of AI unicorns.
Taking SenseTime as an example, in the first half of 2019, 2020 and 2021, its R&D expenses were 1.916 billion yuan, 2.454 billion yuan and 1.772 billion yuan, respectively, an increase of 125%, 28% and 45% year-on-year; in the same period, the revenue growth rate was only 63%, 13.8% and 91.8%, respectively.
A16z, a well-known venture capital institution in Silicon Valley, has mentioned that the gross profit margin of the software service industry is above 60%-80%, while the gross profit margin of most AI companies in the world is usually at the level of 50%-60%.
During the reporting period, SenseTime's gross margin increased from 56.5% in 2018 to 56.8% in 2019 and further improved to 70.6% in 2020 and 73.0% in the first half of 2021.
SenseTime is already one of the best GROSS MARGIN performers among the four AI dragons. However, as an "artificial intelligence software company", SenseTime's gross profit margin is still relatively low in the software service industry.
However, with the mass production of SenseTime's AI models, the AI models and applications are deployed and reused in a variety of scenarios through the software platform, and the software products are generally prioritized in sales, and its gross profit margin is also increasing year by year.
SenseTime's revenue mainly comes from smart business, smart city, smart life and smart cars. At present, the first two contribute almost 85% of SenseTime's revenue.
In fact, most of the customers of ai four dragons are concentrated in ToG and ToB and direction, and such customers usually have a long payment cycle and even have the risk of default.
From 2018 to the first half of 2021, SenseTime's accounts receivable reached RMB1.332 billion, RMB2.615 billion, RMB3.748 billion and RMB3.926 billion, accounting for 71%, 86%, 108.7% and 113.9% of revenue respectively. As of the first half of this year, senseTime's accounts receivable balance for more than one year was close to 1.3 billion yuan.
In addition to its own development needs, SenseTime, which has completed 12 rounds of financing, is increasingly facing the pressure of investors to exit.
A report in Caijing magazine last November said that at that time, some old shareholders began to privately discount the shares of The Soup, "giving a valuation of $9 billion."
Some of SenseTime's institutional shareholders are not financially optimistic. For example, SoftBank, which currently holds 14.88% of its shares in SenseTime, but in Q1 this year, SoftBank Group's net income fell by 61% year-on-year; profit fell 90% year-on-year.
In SoftBank's portfolio in China, although Didi has already IPO, the current market value is less than $40 billion due to the uncertain future of regulatory consolidation; the market value of Shell, Manbang and Shoumen Education has shrunk sharply compared with the IPO.
Financial pressure coupled with the uncertainty brought about by regulatory policies, SoftBank will naturally increase the demand for liquidation of invested companies.
Suning Tesco invested $50 million in SenseTime in 2018, but it was plagued by a financial crisis this year, and founder Zhang Jindong had no choice but to leave.
For the AI Four Dragons, today they have reached the key node of the IPO, and even success or failure in this battle.
For example, Yitu Technology, in June this year, the listing of Yitu Technology Science and Technology Innovation Board was terminated, financing channels were blocked, capital pressure was doubled, and it was exposed to significant layoffs.
In order to alleviate the financial pressure, Yitu broke his arm to survive and sold his medical business. Cloud, where the IPO is progressing slowly, has also exposed rumors of salary cuts from before.
<h2>The Internet's "only fast and not broken" does not work</h2>
Only fast is not broken, big fish eat small fish, exchange losses for scale, crush opponents with financing advantages and eventually form a monopoly position, these are the tried and tested ways to play in the field of Internet venture capital.
From JD.com to Pinduoduo, from Meituan to Didi, from Full Gang to Dingdong to buy vegetables, these players who run out are all successful footnotes to this set of playing styles.
Taking Meituan as an example, in the 8th year after its establishment (that is, 2018), it was already the top brand in the field of life services, and by Q1 of 2018, Meituan's share in the catering takeaway market reached 59.1%, far ahead.
In terms of revenue, from 2015 to 2018, Meituan's revenue was 4 billion yuan, 13 billion yuan, 33.9 billion yuan and 65.23 billion yuan, an increase of 223.2%, 161.2% and 92.3% respectively year-on-year.
When the entrepreneurial outlet shifted from business model innovation to technological innovation, many people are still using the original logic.
In recent years, artificial intelligence is sweeping the world with lightning speed. It is not a simple outlet, it is regarded as a new thing of the same magnitude as the steam engine, electricity, and computers, and its emergence means that the prelude to the Fourth Industrial Revolution has begun, and mankind is about to enter a new era.
In the face of such an opportunity of the times, a large number of capital and entrepreneurs have poured into the AI field in recent years, and the crazy financing story has been staged again. But it turns out that the logic of the Internet's "only fast is not broken" has failed.
In the process of "fighting" from thousands of troops, the AI Four Dragons have also experienced multiple rounds of financing, and the amount of gold absorption is huge: SenseTime introduced 12 rounds of financing totaling 5.2 billion US dollars in the 7 years of its establishment; Megvii raised 9 rounds of financing in 7 years, absorbing 7.46 billion yuan; Cloud raised nearly 3 billion yuan from 6 years; and Yitu raised a total of not less than 3 billion yuan in 10 rounds.
However, after 6-10 years of development, the AI Tigers are far from forming an absolute leading position in the market today.
In June this year, IDC released the "Chinese Intelligent Software and Application Market Research Report 2020", showing that the market size of Chinese intelligent software last year was nearly 23.1 billion yuan, of which the market size of computer vision reached 1.66 billion US dollars (about 11.4 billion yuan). Among them, SenseTime ranks first, but the total market share of AI Four Dragons is less than 50%.
In terms of growth rate, ToB's AI companies are much slower than consumer Internet companies.
Judging from the comparable data disclosed, from 2019 to 2020, senseTime's revenue growth rate was 63% and 13.8% year-on-year, respectively; megvii's data in the same period was 47% and 10%. The cloud has seen negative year-on-year revenue growth from even in 2020.
Chart Note: According to the chart, the 2020 revenue only contains data from January to June
In May this year, Huang He, a partner at Northern Lights Venture Capital, mentioned in an interview with Jiazi Lightyear that in the industrial and technological fields, the characteristics of to B are inherently fast; from the perspective of the maturity process of technical products, it is also necessary to follow objective laws, and it is necessary to go through many iterations to be stable and mature. It all takes time, and this natural law doesn't change dramatically by throwing money at it.
<h2>What is the valuation of 100 billion? </h2>
According to Sullivan's report, SenseTime is the largest ARTIFICIAL intelligence software company in Asia and the largest computer vision software company in China in terms of revenue in 2020.
To be more specific, in terms of software-related revenue in 2020, SenseTime is the largest computer vision software company in China for enterprise applications and urban management applications, and the second largest computer vision software company for consumer applications. SenseTime claims to be a leader in autonomous driving technology and partnerships with Chinese and global automotive companies.
SenseTime's founding team comes from the Multimedia Laboratory of Chinese University of Hong Kong, and the core members Are Tang Xiaoou, Xu Li, Wang Xiaogang, Xu Bing, etc. are all scientists in the field of AI. As of June 30 this year, SenseTime has 40 professors leading its R&D efforts and more than 5,000 employees, of which about two-thirds are scientists and engineers, with an average age of 31.
SenseTime was born almost with a golden key, and it is also wildly sought after by the capital market. In its third year of existence (2017), SenseTime was valued at $3 billion; in 2018, it was valued at more than $4.5 billion. To this day, it continues to maintain its position as the world's largest total financing and the highest valuation of artificial intelligence unicorns.
At the beginning of 2019, SenseTime co-founder Xu Bing revealed that according to the scale of revenue for the year, including future forecasts for revenue, they will soon be able to achieve a valuation of tens of billions of US dollars.
In July 2020, some PE investors revealed that SenseTime was seeking private financing at a pre-investment valuation of $10 billion.
According to the latest report, SenseTime's current valuation has reached US$12 billion, and it will raise at least US$2 billion to list in Hong Kong, according to this rough calculation, SenseTime's valuation will exceed 100 billion yuan.
If you calculate according to SenseTime's revenue of 3.446 billion yuan in 2020, the P/S value corresponding to 100 billion SenseTime is about 30 times.
In A-shares, Hikvision (002415) and Dahua (002236) are SenseTime's peers in ai computer vision applications.
According to the 2020 revenue and current market value, Hikvision and Dahua shares correspond to 7 times and 2.5 times the P/S value, respectively.
SenseTime also has similarities with Baidu in AI, especially in the field of smart cars.
In Hong Kong stocks, Baidu, as the "first AI stock", currently has a market value of only HK$429.4 billion (about US$55.1 billion), excluding Baidu's US$26 billion in cash, cash equivalents and short-term investments.
Baidu's various assets are valued at only about $30 billion.
Baidu's revenue scale in 2020 was 107.1 billion yuan, of which AI innovation business (including intelligent cloud, Xiaodu speaker, Apollo, etc.) revenue was 12.4 billion yuan; for the whole of last year, Baidu achieved a net profit (non-US GAAP) of 22 billion yuan.
In 2020, SenseTime's revenue of 3.446 billion yuan is less than 30% of Baidu's AI innovation business revenue, and it is less than a fraction of Baidu's annual revenue of 107.1 billion yuan.
But now SenseTime's valuation of more than $10 billion is about equal to 1/5 or even 1/3 Baidu.
Is SenseTime overrated, or is Baidu undervalued?
Huang He, a partner at Northern Light Ventures, once warned against the Internetization of industrial and technological investment. He believes that high valuation should correspond to high growth, ignoring the law of development and the basic criteria of valuation in this field, not only can not make a profit on investment, but also may boost the industry bubble, and even cause a vicious circle.
There is no doubt that artificial intelligence as the infrastructure of all walks of life in the future, the development prospects are promising.
Sullivan's report argues that AI software will be one of the fastest-growing business segments over the next decade. In 2025, the global artificial intelligence software market size will reach $121.8 billion, with a compound annual growth rate of 31.9% from 2020.