With the birth of the first AI stock and the market value of SenseTime exceeding HK$180 billion, the industry may enter a new stage of development

Produced | Sanyan Finance

On December 30, on the penultimate day of 2021, SenseTime, known as the "first ai stock", officially landed on the Hong Kong stock market.

On the first day of listing, the stock price rose sharply, with the highest intraday increase of 22.08%. At the close of the day, SenseTime's share price rose 7.27%, with a total market value of nearly HK$140 billion.

SenseTime opened today with another surge, closing up 33% at noon and having a market capitalization of more than HK$180 billion.

In fact, as early as 2019, Chinese AI startups have begun to compete for the title of "AI first stock", but two years have passed, stop-and-go, and thorns are dense.

Similarly, SenseTime's IPO road is full of difficulties and obstacles. In August this year, SenseTime submitted an application for listing on the Hong Kong stock market. On November 22, SenseTime passed the hearing of the Hong Kong Stock Exchange.

However, on the day of SenseTime's original planned Hong Kong listing pricing, the U.S. Treasury Department included SenseTime in the investment blacklist of "China's military-industrial complex enterprises", and SenseTime had to postpone the IPO plan. However, only a week later, the restructured SenseTime launched a global offering again and successfully listed against many difficulties.

"There is a continuous large-scale investment in technology to bring innovation, and then there is business model development, which is a rare road." The hard part is the uncertainty of the business. Xu Li, co-founder, chairman and CEO of SenseTime, said in his listing speech on December 30.

SenseTime opened up the capital closed loop and brought a glimmer of light to the entire industry. In the long run, the artificial intelligence industry may enter a new stage of development.

"The decision to change the world can take just 7 seconds"

SenseTime is a technology company that came out of the "laboratory". In 2014, Professor Xiaoou Tang, who teaches in the Department of Information Engineering at the University of Chinese, Hong Kong, founded SenseTime.

Co-founders Xu Li, Wang Xiaogang and others came out of this laboratory.

The typical feature of AI startups is that the genesis team is usually a technology master and has a deep academic atmosphere.

Megvii was founded in 2011 by three founders, Yin Qi, Tang Wenbin and Yang Mu. They all graduated from the "Yao Class" of Tsinghua University. Yuncong was incubated from the Chinese Academy of Sciences and is known as the "AI National Team" in the industry. YITU Technology was co-founded by Zhu Long, a researcher at New York University, and Lin Chenxi, then technical director of Alibaba Cloud.

As of June 30, 2021, the company has 40 professors leading research work and 3593 technical research and development personnel, including more than 250 doctoral and doctoral candidates.

At yesterday's listing ceremony, Dr. Xu Li recalled the entrepreneurial scene of that year.

Xu Li said, "It is difficult to imagine a group of graduates of famous schools who chose to join a startup company with a low salary, not even a good name, and even a commercial and residential building in the office."

He jokingly said, "If there's anything particularly appealing, it's probably 7 seconds from bed to desk." ”

In the eyes of this group of technical men from famous schools, there is nothing more exciting than changing the world, "Why not personally promote the landing of technology to achieve our dreams." And the decision to change the world may only take those 7 seconds."

The mission of SenseTime is to adhere to originality and let AI lead human progress.

However, original use may be more academically concerned, but original technology in the commercial field does not necessarily have a first-mover advantage.

Especially in the AI industry, many technological innovations lack practical application scenarios, or there are market exploration and education periods.

So, how to understand SenseTime's "sticking to the original" business model?

In this regard, Xu Li has previously explained in detail. He pointed out that there are actually three meanings in this:

1. Be a long-termist and insist on investing in originality;

2. The intensity of investment innovation should be large enough to achieve "escape speed";

3. From defining technology to defining the industry, be a leader.

Xu Li said that these three layers of meaning are progressive. Its practical significance lies in the guidance of strategy, often resource allocation and business trade-offs. "The long-term investment in research and development, the construction of artificial intelligence infrastructure, and the layout of computing power platforms are closely related to our mission. Some of our businesses, judging from the traditional business logic, are not yet successful, and even some of them are failures. But they do redefine, or change, even a small part of the way industries and lives are lived, and that should be recognized by us."

From Xu Li's expression, it can be seen that SenseTime does not focus on the rapid realization of business, what it wants to do is to redefine and change an industry or life.

In fact, the vast majority of AI startups are facing the problem of commercial monetization, which is fundamentally determined by the industrial characteristics of artificial intelligence: first there is sustained large-scale investment, and then there is a business model.

A few years ago, under the AI outlet, AI companies received a lot of financing.

According to data, in the 10 years from October 2011 to the present, China's AI industry has received a total of 216.9 billion yuan in financing, 97% of which occurred after AlphaGo defeated the World Champion of Go, and investors include Sequoia China, Hillhouse Venture Capital, Qiming Venture Capital, Zhen Fund, etc.

Prior to the listing, SenseTime completed a total of 12 rounds of financing, with a total financing amount of US$5.2 billion (about 33.658 billion yuan).

However, in the past two years, the financing events in the AI industry have decreased significantly, and the listing is also difficult, and many AI companies have continued to spread the news of layoffs and salary cuts.

In this context, we can also understand why the listing of SenseTime has caused such a big sensation.

For a long time, the high investment in research and development and the difficulty of hematopoiesis in the AI industry have led to serious losses for a long time.

AI startups are desperate to get the funds they need for long-term growth through listing, and SenseTime's successful breakthrough means that the capital market's confidence in AI startups has been restored.

In addition, from the perspective of financial data, the composition structure of SenseTime's research and development, revenue and loss is constantly being optimized, and AI startups may enter a stage of rapid growth.

SenseTime's financial structure is optimized

Revenue has entered a stage of large-scale growth

Research and development continues to grow

Losses are narrowing

The AI industry continues to invest heavily in research and development as the biggest reason for losses.

As a leader in domestic AI companies, SenseTime far surpasses other companies in terms of revenue and R&D investment.

Taking 2020 as an example, SenseTime's revenue of 3.45 billion yuan exceeded the sum of 2.5 billion yuan of Yuncong, Megvii and Yitu.

SenseTime in 2018, 2019, 2020 and the first half of 2021, the revenue was 1.85 billion yuan, 3.03 billion yuan, 3.45 billion yuan and 1.652 billion yuan, respectively, and SenseTime's revenue has entered a stage of large-scale growth, with a year-on-year increase of 91.87% in the latest half year.

In terms of R&D investment, 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, SenseTime invested 1.772 billion yuan in research and development, which was 100 million yuan higher than the current revenue. In the past three and a half years, SenseTime has invested nearly 7 billion yuan in R&D.

The good news is that during the above reporting period, SenseTime's corresponding net losses were 150 million yuan, 1.16 billion yuan, 710 million yuan and 580 million yuan, respectively, which was narrowing year-on-year.

In addition, SenseTime's gross profit margin is now at a high level in the industry, currently maintaining above 70%.

The AI industry needs long-termists

SenseTime is building infrastructure for the AI era

Since its inception, SenseTime has always invested in original technology research and achieved many industry firsts.

For example, the first to achieve the accuracy of AI visual algorithms to exceed people; the first to achieve subway brush faceless payment in tens of millions of population cities; in the application of smart cities, the first to put forward the concept of "city-level search engine" and "long-tail scene", which has now become the standard customer demand of hundreds of cities.

As early as 2017, SenseTime achieved the first pure visual perception driverless test in a controlled scene on a rainy day, becoming the first Chinese company to export China's original technology to top overseas car manufacturers.

In the field of smart healthcare, SenseTime was the first to develop AI software that has obtained CE certification for EU medical devices. It is also the first to launch a full set of teaching materials for artificial intelligence primary schools, junior high schools and high schools, opening a new era of intelligent basic education for Chinese engineering.

Since its inception in 2014, SenseTime has won more than 70 global championships, more than 600 top academic papers, more than 8,000 AI invention patents, and ranked first in the world in the total number of papers published at the world's three major computer vision conferences.

It is precisely under the support of original technology research that SenseTime's full-stack artificial intelligence capabilities have been continuously enhanced, covering key technology areas such as perception intelligence, decision intelligence, intelligent content generation and intelligent content enhancement, as well as key capabilities including AI chips, AI sensors and AI computing infrastructure.

At present, SenseTime's business covers four major sectors: smart business, smart city, smart life and smart car.

Artificial intelligence technology is gradually evolving from the original single direction, single route, and gradually to a general technology of large-scale parameters.

At present, AI has been quite perfect in some head needs, but long-tail application scenarios, such as garbage dumping, manure truck emissions, bicycle illegal parking, etc., are still difficult to achieve complete coverage.

In this context, SenseTime has built a new type of artificial intelligence infrastructure - SenseCore AI device.

SenseCore SenseTime AI device can be seen as the integrator of SenseTime's most advanced algorithms, computing power and systems. It is composed of three parts: the computing power layer, the platform layer and the algorithm layer.

SenseCore can promote the research and development of general-purpose technologies, comprehensively solve the needs of long-tail applications, and achieve high-efficiency, low-cost and large-scale AI innovation and empowerment by significantly reducing the cost of artificial intelligence production factors.

To understand it simply, similar to our search engine through the Internet, SenseCore is to build a digital search engine and recommendation system in the physical space, accelerate the landing of AI, and deliver AI applications to various fields and scenarios in urban management, enterprise services and personal life.

According to the prospectus, 60% of the funds raised by SenseTime were used for R&D investment to enhance R&D capabilities.

Among them, 10% of the AI infrastructure SenseCore is used for supercomputing centers, 10% for AI chips, and 15% for model capability improvement. A further 25% will be invested in technical models, products and industry-university-research.

Currently, SenseTime is the number one revenue company in Asia in the AI industry. It also has a place in the global AI industry, and there may be more leading opportunities in the future.

For AI startups, continuous scientific research investment is poison and antidote, and more important is patience and direction behind it.

There is never a shortage of trendsetters in the market, but there are not many who can cultivate deeply in an industry without being affected by immediate interests.

The listing of SenseTime has brought a glimmer of light to the industry, but how many can be like SenseTime?

The moment of the AI industry's big explosion may have to wait, so wait another 5 years to see what kind of world this wave of trends can create.