IPO approved? ! How does SenseTime get rid of the common predicament of AI companies with the entry of Metaverse and autonomous driving?
Meta Cosmos Autonomous Driving
Recently, some media reported that SenseTime, one of the "AI Four Little Dragons", was approved for an IPO in Hong Kong stocks and plans to raise US$1 billion. However, Shangtang did not respond positively to this rumor, and the relevant person in charge said to the media that he would not comment on this.
Earlier news revealed that SenseTime may list A+H. SenseTime submitted a Hong Kong stock listing application on August 27 this year, but this is actually the latest company among the "AI Four Dragons" to submit a prospectus. Megvii submitted the prospectus to the Hong Kong Stock Exchange as early as August 2019, but there has been no follow-up; until March 2021, the Science and Technology Innovation Board disclosed the news of Megvii’s submission of the IPO application and completed the meeting in September. One step away from listing.
Yitu Technology was the first of the AI four dragons to submit a prospectus to the Sci-tech Innovation Board, and one month later, in December 2020, Yuncong Technology also submitted a prospectus to the Sci-Tech Innovation Board. However, on June 30 this year, Yitu took the initiative to withdraw its application for listing on the Science and Technology Innovation Board; but after the successful meeting of Yuncong in July, it was only a matter of time before it went public.
AI companies have gathered together to seek listing in the past year. From the prospectus of these companies, we can find some difficulties in the domestic AI industry in recent years.
SenseTime is not immune to the common dilemma of AI companies
The AI company's loss has been discussed enough. After several years of development, the commercial landing process of AI companies is still unsatisfactory. After several years of AI craze, the market has begun to calm down. After the bubble on the AI track has gradually faded, it is now increasingly difficult for companies to raise funds.
Source: SenseTime's prospectus
According to Shangtang's prospectus, revenues in 2018, 2019, and 2020 were 1.853 billion yuan, 3.026 billion yuan, and 3.446 billion yuan, respectively, while revenue in the first half of 2021 reached 1.65 billion yuan. During this period, the gross profit margin is still increasing. From 2018 to 2020, the gross profit margin is 56.5%, 56.8%, and 70.6%, respectively, and the gross profit margin in the first half of 2021 has reached 73%.
Even if the gross profit is as high as 70%, Shangtang still has not escaped the fate of huge losses. In 2018, 2019, and 2020, SenseTime lost 3.433 billion yuan, 4.968 billion yuan, and 12.158 billion yuan respectively. In the first half of 2021, it also lost 3.713 billion yuan.
Compared with the data of several other AI companies, taking Kuang as an example, its revenue in 2018, 2019, and 2020 are respectively 854 million yuan, 1.260 billion yuan, and 1.391 billion yuan, and net losses are 2.8 billion yuan and 66.39 respectively. Billion yuan, 3.327 billion yuan. In the first half of this year, although Megvii ushered in a year-on-year increase in revenue, the loss still showed no signs of narrowing: revenue in the first half of the year was 669 million yuan, a year-on-year increase of 91.27%, but the net loss was as high as 1.858 billion yuan.
In the prospectus, SenseTime expressed its belief that future revenue growth depends on the ability to develop new technologies, enhance customer experience, formulate effective commercialization strategies, effective and successful competition, and develop new products and services. But they also said that it is expected that costs and expenses will increase in the future as the company continues to expand its business and operations and invest in research and development, general artificial intelligence infrastructure and geographic expansion.
The trend of AI technology homogeneity is actually forcing AI companies to invest more resources to maintain their advantages, which can be reflected in R&D investment.
SenseTime’s R&D expenditures from 2018 to 2020 were 849 million yuan, 1.916 billion yuan, and 2.454 billion yuan, respectively. R&D expenditures accounted for 45.9%, 63.3%, and 71.3% of revenue. The trend of increasing proportion of revenue. In the first half of 2021, SenseTime's R&D expenditure reached 1.772 billion yuan, which was even 7.3% higher than its revenue during the same period.
However, since its establishment in 2014, SenseTime has completed a total of 12 rounds of financing, with a total amount of US$5.2 billion. This is the confidence that SenseTime can continue to "burn money". According to the data in the prospectus, SenseTime currently has sufficient cash flow. As of June 30, 2021, the company's cash and cash equivalents amounted to 8.926 billion yuan.
Regarding the US$1 billion in funds raised, SenseTime stated in the prospectus that 60% will be used to improve research and development capabilities, including expanding the company’s AIDC computing power, strengthening AI chip design capabilities, and developing free AI chip solutions. AI model related capabilities; 15% for business development; 15% for seeking strategic investment and acquisition opportunities; 10% for operating capital and general corporate purposes.
However, in the face of potentially higher R&D expenditures in the future, going public is the best financing channel for AI companies that are currently cooling down.
AI vision market share is leading domestically. Will autonomous driving and meta-universe be the growth points in the future?
According to the Sullivan report, in terms of revenue, SenseTime is the largest artificial intelligence software supplier in Asia and the largest computer vision software supplier in China in 2020.
From the perspective of revenue structure, SenseTime’s main business can be divided into smart business, smart city, smart life, and smart cars. Among them, smart business and smart city are SenseTime's current main sources of revenue, and the two businesses together account for about 80% of total revenue.
From the perspective of business revenue trends, the proportion of smart city revenue has been increasing in recent years, from 28.6% in 2018 to 47.6% in January-June 2021. The proportion of smart living in total revenue has declined significantly. The data in 2018 was 17.8%, and by the first six months of 2021, this figure had dropped to 8.9%. The revenue of the smart car business has grown steadily, but with the current gradual increase in the demand for smart cars, it has the opportunity to become a new growth point in the future.
In 2017, SenseTime became a strategic partner of Honda, providing it with technologies related to autonomous driving. As of June 30, 2021, SenseTime has accumulated more than 30 domestic and foreign auto companies and has been selected as a supplier of more than 50 models. SenseTime said that in the next few years, they will supply more than 20 million cars with its "Juying Smart Car Platform" products.
In addition to smart cars, SenseTime also has a layout in the recent popular meta-universe concept.
The prospectus shows that the SenseMARS software platform of SenseTime is the technology empowerment platform of Yuan Universe, which is equipped with perception intelligence, decision intelligence, intelligent content generation (including AR/MR and software agents) and other infrastructure (such as cloud engines, etc.) And other functions, and cooperate with mobile phone manufacturers, semiconductor manufacturers, mobile application developers and game developers to jointly build a multi-layer infrastructure.
Currently through the SenseMARS platform, SenseTime has already had many application landing cases: The Forbidden City collections are presented online through AR; the three-body offline immersive experience hall is created; the large-scale offline carnival Bilibili World 2021 virtual experience space; Hangzhou West Lake immersive AR Guided tour; a supplier of intelligent vision technology for the 2022 Asian Games, providing a variety of AI and MR applications.
The prospectus revealed that in 2018, a company’s meta-universe application used a virtual avatar based on SenseMARS.
Of course, for the meta universe, there are various interpretations and visions for the future. But the reality is that Meta Universe has not yet implemented a product in the true sense, and the industry has not yet a clear definition of Meta Universe. The important thing is that for the concept of Metaverse, how long the status quo of the capital carnival can last is still a question. Perhaps Metaverse is just like the AI industry in previous years?