As the unicorn company with the most financing and the highest valuation in the AI field, SenseTime has the greatest opportunity and is also under the greatest pressure.
SenseTime is currently the last company in the "AI Four Little Dragons" that has no clear public listing plan.
A SenseTime technology financing plan recently obtained by a reporter from Caijing mentioned that SenseTime expects to be listed in both A+H shares within three years. An investor close to SenseTime told a reporter from Caijing that he is currently in contact with brokers. As for the specific listing situation, the relevant person in charge of SenseTime told the reporter of Caijing that he would not comment.
The investor revealed to a reporter from Caijing that SenseTime had just completed a round of financing this year, with a pre-investment valuation of US$11.3 billion, but this round of financing "is not a formal round of financing, and it will be completed after receiving the money."
SenseTime's last round of financing was in November 2018, with a joint investment of US$580 million by SoftBank Vision Fund, Fidelity International, and Mirae Asset. This round of financing was already the sixth round of SenseTime's financing. At that time, rumors about SenseTime's IPO in Hong Kong stocks had already appeared.
Up to now, SenseTime has completed a total of about 4 billion US dollars in financing, which is the one that "attracts the most money" among the four AI dragons. Among the other three companies, Megvii's total financing is approximately US$1.35 billion, Yitu is over US$400 million, and Yuncong is approximately RMB 3.5 billion. Both Megvii and Yitu have issued prospectuses, and Yuncong is also conducting IPO counselling.
It can be seen from the prospectus of Yitu and Megvii that the current common characteristics of AI startups are: high valuations, fast revenue growth, but low gross profit, and high debt-to-asset ratio, even if they have received high financing, Still facing financial pressure.
According to the financing plan data obtained by a reporter from Caijing, SenseTime's 2019 operating income was 5.06 billion yuan, with a gross profit margin of 43%, of which smart city revenue was 1.664 billion yuan, with a gross profit margin of 57.8%; smart business revenue was 2.681 billion yuan, with a gross profit margin. 23.9%. The relevant person in charge of SenseTime did not confirm to the reporter of Caijing whether this data is accurate. He responded that: “Financial data cannot be disclosed to the outside world at present.”
Before this, many people in the industry believed that Shangtang might become the "first AI stock." Business information shows that SenseTime was established in 2014, 3 years later than Megvii Technology. Intensive financing mainly occurred in 2017-2018. For investors, SenseTime is not in a hurry to go public.
However, valuation shrinkage has already appeared. An investor close to SenseTime told Caijing that SenseTime’s financing this year was a bit difficult because the previous financing speed was too fast and the valuation was too high. Many investors believe that now It's not worth investing in. In addition, the investor also mentioned that some old shareholders started to trade Shangtang's shares privately, but it was difficult to sell them. "The current valuation is $9 billion."
01 Common problems with low gross profit margin
Usually company financing is based on two purposes, one is lack of money; the other is that more funds are needed for development and expansion, but the existing business cannot support it.
The data in the financing plan shows that SenseTime's revenue in 2019 was 5.06 billion yuan, gross profit margin was 43%, gross profit was 2.16 billion yuan, and net profit and net profit margin were not disclosed. However, it disclosed the net interest rates for 2017 and 2018, which were 2% and 3%, respectively, and net profits were 11 million and 59 million, respectively.
An investor close to SenseTime told the Caijing reporter that SenseTime was not profitable in 2019, which is also one of the factors that affected SenseTime's financing this year.
The gross profit margin of 43% is low among the leading AI companies. Megvii's gross profit margin in the first half of 2019 was 64.6%, and Yitu's gross profit margin in 2019 was 63.9%.
Many investors who are concerned about the AI field have expressed their concerns about the low gross profit of AI companies to Caijing reporters. Xiong Weiming, a partner of Huachuang Capital, once mentioned, “This is because they are mainly integrated in the industry. Lower gross profit margins.” The aforementioned investors familiar with SenseTime said that gross profit cannot rise mainly because of two factors: a large proportion of hardware and a heavy service.
The landing projects in the AI field include a large number of non-standardized products and services, and there are obvious differences between each project, requiring a lot of effort to understand needs, field inspections, procurement and services.
SenseTime’s business is divided into smart city, smart business and other businesses, and other businesses include education and medical care. In 2019, the main revenue contribution is smart business, with revenue of RMB 2.681 billion, accounting for 53% of total revenue. This revenue figure has increased by 236% compared to 2018, but the gross profit margin has dropped from 51.2% to 23.9%.
In the business matrix of SenseTime, the business scope of smart business includes smart Internet of Things, smart traffic, identity verification and precise advertising. It is mentioned in the plan that the gross profit margin of this business has decreased mainly because many projects in the intelligent Internet of Things are in the infrastructure construction stage. In addition, in order to promote the sale of face recognition equipment, regional distributors are used to serve small-scale customers. Dealers make profits in exchange for channels and users.
In the industry cultivation stage, companies usually sacrifice some short-term benefits in order to obtain users, in order to open channels, establish brand awareness, and lay the foundation for subsequent rapid growth. However, many AI industry practitioners mentioned that the current competition among the AI Four Dragons is fierce, and the price war will not stop in the short term, and there is no significant difference in the capabilities of basic products such as face recognition.
The gross profit of more than 40% is not low in many other industries. However, AI is a high-tech industry, and AI companies are facing long-term high-tech investment, the project implementation cycle is long, and there is pressure to pay back. Such gross profit margin will appear to be not high enough.
The plan mentioned that with the continuous implementation of multiple projects in the future, the gross profit rate can be increased to more than 50%, but SenseTime is not only competing with the AI Four Dragons, but also information companies in various industries and comprehensive strengths. Stronger tech giants.
02The battle for smart cities
The rapid growth of smart city revenue is also one of the business areas where SenseTime can flex its muscles. This prospectus takes smart cities as the main source of income for SenseTime in the future. In 2019, SenseTime completed a total of 1.664 billion yuan in revenue in the field of smart cities, a year-on-year increase of 147%, and its revenue accounted for 32.9%. Management predicts that in 2020, smart city revenue will reach 4 billion yuan, accounting for 42.7% of the revenue.
However, smart cities are also key businesses of Megvii and Yitu. These leading AI companies once regarded security as their main business, and now everyone has included security in the big basket of smart cities.
Not only are AI giants, smart cities are currently the most potential digital market, and they are also a battleground for technology companies. According to data from the consulting agency IDC, in 2019, China's smart city spending is 20 billion U.S. dollars, and the annual growth rate is expected to exceed 15%.
At present, SenseTime’s smart city revenue is highly concentrated. In 2019, the top five regions accounted for 74.3% of revenue. They are Shenzhen (389 million), Shanghai (263 million), Shandong (260 million), and Yunnan (195 million). ) And Guangdong (130 million).
Star Capital founding partner Yang Han Song told the "Financial" reporter, want to do good Chi Hui-city business, the core competitiveness is the ability to integrate resources, smart city involves a lot of supporting industries, real estate, finance and industry, AI techniques The proportion in the middle is not high. However, according to the data in the plan of SenseTime, the visual analysis part of smart city projects usually accounts for 10%-30%.
Shangtang has settled in many cities. In July 2020, SenseTime settled in Xi'an "SenseTime Xi'an Research Institute"; in December 2019, SenseTime established a regional headquarters in Qingdao; in August 2019, SenseTime announced that its China headquarters and global R&D headquarters will be located in Shanghai; 2018 SenseTime’s third regional headquarters was settled in Hangzhou in October, and the first and second regional headquarters were in Beijing and Shenzhen respectively; in May 2018, SenseTime registered and established a wholly-owned subsidiary in Chengdu Tianfu New District; in November 2017, Tang established SenseTime Southwest Headquarters and Southwest Research Institute in Chongqing. In January this year, the "Hunan Daily" reported that the docking between Changsha and Shangtang Technology is also underway.
It is mentioned in the plan that in 2018, SenseTime's smart city business has covered 78 cities, and it will increase to 127 in 2019. In the future, it plans to focus on developing second- and third-tier cities.
Some investors worry that to settle in multiple cities, laboratories and project teams need to be established locally, and personnel management may be problematic. But this can indeed quickly establish contact with local governments.
The smart city project sounds beautiful. The amount of smart transformation of a city is huge and it can quickly gain popularity, but the project cycle is usually more than 3 years. The process includes top-level government design, pilot projects, feasibility studies, project initiation, budget declaration, Detailed plan design, bidding, signing, implementation, installation and deployment, customized development, initial inspection, down payment, final inspection, payment collection, maintenance and after-sales service, etc. It is mentioned in the plan that Shangtang mainly participates in projects in phases at present, and is still working hard to increase its share in the project.
Head companies are more likely to gain a higher share of government projects. SenseTime is already well-known in the AI field, but the other three companies are also well-known, have government relations, and have technical strength. Moreover, it is not only the AI unicorns who have AI capabilities, their common competitors come from different tracks, and they are all very powerful. They include platform giants such as Ali, Huawei, Tencent, and Baidu , as well as IT giants.
03 Capital pressure and valuation bubble
Comparing the financial data of Shangtang, Megvii and Yitu, it can be found that the same problem faced by these three leading AI companies is that they have high debt-to-asset ratios. The prospectus shows that the asset-liability ratio in the first half of 2019 is 218.79%, and the asset-liability ratio in 2020 is 252.28%. In the financing plan, SenseTime's asset-liability ratio in 2019 is 102.35%.
Comparable listed companies, the Hikvision 2020 asset-liability ratio was 38.49%, ArcSoft is 10.64%, IFLYTEK is 42.33%.
In addition, at the end of 2019, SenseTime's inventory increased by 304% year-on-year, and there is also a certain amount of pressure to pay back. The plan mentioned that the payback period for government projects is about 348 days.
The high debt-to-asset ratio makes these AI unicorns feel the financial pressure. The aforementioned investors mentioned that SenseTime currently has about 20 billion yuan in cash on its accounts, which can support the listing in 2 years, but there is still a valuation. The problem is too high.
According to the financing plan, SenseTime’s current valuation has returned to a reasonable level due to revenue growth. The current P/S (market-to-sales ratio) multiple is 11.5. In comparison, it is about 30 times as shown in the figure. Is 13.3 times.
Usually when the company fails to achieve profitability , P/S is used for valuation. However, there are certain doubts about the income quality of AI companies.
SenseTime’s financing plan mentions smart education business. SenseTime first promotes smart education products and services to local education bureaus at the provincial and municipal levels, and then further promoted to the school level by the local education bureaus. Related sales and service contracts are passed through integrators. Sign and collect payment for products and services through integrators. On average, each student receives a fixed fee of 600-800 yuan per year, and a certain discount is provided based on the total contract amount.
Products and services mainly include educational platforms and teaching materials. The AI textbooks displayed on Shangtang's official website are "Introduction to Artificial Intelligence" and "Basic Artificial Intelligence (High School Edition)". Whether book sales can be counted as AI-related income is still unconclusive.
Many investors mentioned to the Caijing reporter that AI companies are currently striving to expand their revenues in order to go public, but many of these revenues are not directly related to AI technology. Many times they use the concept of AI, or Do some basic data analysis.
Yang Ge mentioned that the current evaluation of AI startups should only look at the revenues related to AI. "Currently, corporate customers and government customers don’t have much demand for AI. In a project, how much revenue comes from AI? Coming from outside of AI, this is where the real strength differences of AI companies are reflected."
A large amount of capital has entered, allowing AI unicorns such as SenseTime to grow rapidly, but this momentum seems to have stopped, and investment institutions dare not to overestimate AI companies. Data from data service providers’ business cards shows that in 2020, the AI sector’s financing enthusiasm continues to decline, and the amount of financing for a single project has declined significantly. Up to now, a total of 305 financings have been completed, with a total amount of approximately 24.33 billion yuan. In 2018, these two figures They were 523 transactions and 66.71 billion yuan respectively.
The valuation bubble in the AI field is being broken. AI companies are trying to ease the pressure on valuation by expanding revenue, but revenue is not made out of thin air, and market demand is still in a slow growth stage.
From face recognition to security, to smart cities, Internet of Things, and chips, the AI four dragons have begun to choose different development paths. There are human-based platform development, some are deep in the industry, and some open up new battlefields, but the core is still based on visual recognition. algorithm. At present, Yitu, Megvii, and Yuncong have all entered the listing process. The listing can broaden financing channels, reduce financing costs, and ease capital pressure. However, given various factors, the road to IPO of AI technology companies does not seem to be smooth. The technology IPO has lasted for more than a year, and there is still no substantial progress.
In the past few years of competition, the field of visual recognition has formed the AI four dragons pattern, but a small track is difficult to support 4 companies with a valuation of several billion or even tens of billions of dollars. According to data from the Foresight Industry Research Institute, in 2019, the size of China's machine vision market is 6.55 billion yuan. Next, the competitive landscape in the AI field will usher in new changes. Who will be out, who will be acquired, and who can win-take-all? As the AI company with the highest valuation, SenseTime faces the greatest opportunity, and perhaps the greatest pressure that comes with it.
No comments:
Post a Comment