Produced by Radar Finance and | Li Yihui, ed. | Deep Sea
The Hong Kong Stock Exchange disclosed on February 23 that Beijing Fourth Paradigm Intelligent Technology Co., Ltd. updated its prospectus to the main board of the Hong Kong Stock Exchange, with Goldman Sachs and CICC as its joint sponsors.
On August 13 last year, The Fourth Paradigm submitted a prospectus in Hong Kong stocks. Unfortunately, due to the failure to pass the hearing within 6 months, on February 14 this year, the status of the IPO application of the fourth paradigm has turned "invalid".
Now that the second table is handed over, the financial situation of the fourth paradigm has not changed much, and the loss margin has gradually expanded. According to the prospectus, from 2018 to the end of September 2021, its corresponding net losses were 372 million yuan, 718 million yuan, 750 million yuan and 1.478 billion yuan, respectively, and the cumulative loss has reached 3.3 billion yuan.
From the perspective of the company's financial performance, rising losses and insufficient hematopoietic capacity still plague the fourth paradigm. Many analysts point out that improving the level of technology application and opening up profit growth points are still the top priority of AI companies, and simply financing and listing cannot solve the profit dilemma of enterprises.
Founded 7 years ago, financing nearly 10 billion
Before starting a business, Dai Wenyuan, the founder of the Fourth Paradigm, was labeled as a scholar and a genius scientist.
According to the data, in 2005, while still studying at Shanghai Jiao Tong University, Dai Wenyuan's team won the world championship of the ACM International College Student Programming Competition.
At that time, the Department of Computer Science of Shanghai Jiao Tong University was not equipped with artificial intelligence-related teachers, and the department recommended Dai Wenyuan as a visiting student to the Hong Kong University of Science and Technology, under the tutelage of Yang Qiang, professor of the Department of Computer Science and the first academician of the International AI Association in the Chinese community. Since then, Dai Wenyuan has stepped into the field of AI.
In 2009, Dai Wenyuan joined Baidu as a scientist as the chief architect of Baidu's most important Fengchao strategy, and as a leader in Baidu's machine learning, he participated in and directed Baidu-related searches, Baidu Brain and other products. During the Baidu period, Dai Wenyuan rose to the T10 level in three years, and has maintained this record unbroken, known as "Dai Shen".
Soon after, Dai Wenyuan worked as a chief scientist at Huawei's Noah's Ark Laboratory and won huawei's "President Award" president award.
In September 2014, Dai Wenyuan turned around and joined the entrepreneurial army, joining three friends Yang Qiang, Chen Yuqiang and Hu Shiwei to create the fourth paradigm in a rented house in Futian, Shenzhen.
At first, the fourth paradigm tried to make an efficient tool-level product, but eventually it was dispersed due to customer needs. In July 2016, the team developed Prophet, a machine learning platform that can be used by non-professionals.
In the direction of segmentation, unlike the artificial intelligence startups in the same period, which aimed at the fields of vision and speech, the fourth paradigm focuses on the field of decision-making AI and provides platform-centric artificial intelligence solutions.
Perhaps because it caught up with the boom in the development of AI, at the beginning of its establishment, the fourth paradigm was quite popular in the capital market. According to Tianyan, in October 2015, sequoia fund invested in the angel round of the fourth paradigm. Subsequently, Innovation Factory, Bank of China, Legend Holdings, Primavera Capital and other institutions entered the fourth paradigm investor list.
As of the last round of strategic investment from Tencent Investment in February last year, the fourth paradigm has received a total of 9 rounds of financing, with a cumulative amount of nearly 10 billion yuan and a valuation of nearly 3 billion US dollars.
As of the IPO, Dai Wenyuan, the founder of the fourth paradigm, held 106 million shares, holding 24.25%. In addition, Paradigm Investment holds 14.61% of the shares, Sequoia Capital holds 7.37%, Boyu Jingtai holds 3.23%, Guoxin Tuss holds 2.77%, and Sino One holds 2.76%.
At present, The Fourth Paradigm has developed end-to-end enterprise-level AI solutions, including optional supporting infrastructure, operating systems, AI developer kits and applications to meet the needs of enterprises in the ai construction process for applications, platforms and infrastructure.
Among them, the Prophet platform is the backbone of the company's solution, which allows enterprises to build tailor-made artificial intelligence systems. These systems are driven by automated machine learning algorithms of the fourth paradigm, automating the processes of machine learning, application, decision-making, and evaluation, and are characterized by supporting rapid and easy modeling, providing low- or no-code development environments, and requiring a high degree of involvement without AI experts.
In addition, The Fourth Paradigm provides ready-to-use AI applications that users can directly deploy and use to optimize their marketing, risk control, operations, and other business links. In June 2021, the company launched the Enterprise AI App Store, a marketplace for AI apps that provides users with choice.
It lost 3.3 billion yuan in more than three years
The prospectus quoted China Insight Consulting as saying that in terms of revenue in 2020, the fourth paradigm occupies the largest market share in China's platform-centric decision-making artificial intelligence market.
Industries served include, but are not limited to, finance, retail, manufacturing, energy and power, telecommunications and healthcare. In fact, the fourth paradigm is the first start-up to be invested by five state-owned banks, including Industrial and Commercial Bank of China, Agricultural Bank of China, Bank of China, Bank of Communications, and China Construction Bank, whose shareholders are also its main customers.
In 2020, the company served 47 Fortune Global 500 companies and public listed companies (i.e. benchmark users), and in 2019 and 2020, the net income expansion rate of benchmark users was 250% and 167%, respectively.
In the context of expanding the use of the company's solutions, the company's average revenue per benchmark user has been increasing significantly. In 2018, 2019, 2020 and the nine months ended September 30, 2021, it was 3.9 million, 8.3 million, 12.3 million and 12.4 million, respectively.
The Company's customers fall into two broad categories, including direct sign-off customers, which are end users who purchase their solutions directly, and Solution Partner customers, primarily third-party system integrators who integrate the Company's solutions into their products to meet the specific needs of end users.
During the Reporting Period, revenue generated by direct customers accounted for 19%, 21%, 15%, 13% and 25% of the Company's total revenue, respectively, and revenue generated from Solution Partner customers accounted for 81%, 79%, 85%, 87% and 75% of the Company's total revenue, respectively.
This revenue structure also illustrates that, in a sense, the fourth paradigm is increasingly like an enterprise solution provider.
The revenue generated from these customers mainly comes from the Prophet platform and products of the Fourth Paradigm, as well as application development and other services. With the accumulation of the number of users and the rise of the average income of users, the revenue of the fourth paradigm has grown rapidly.
Specifically, the Company's revenue increased by 259.7% from $128 million in 2018 to $460 million in 2019 and further by 105.0% to $942 million in 2020. Revenue in the first three quarters of 2021 increased by 134.3% year-on-year to RMB1.345 billion, far exceeding the level of the whole year of 2020.
Despite the strong revenue growth rate, the "curse" of losses criticized by the market still plagues the fourth paradigm.
According to the prospectus, in 2018, 2019 and 2020 and as of September 30, 2021, the fourth paradigm generated net losses of 372 million yuan, 718 million yuan, 750 million yuan and 1.478 billion yuan respectively, and the cumulative loss has reached 3.3 billion yuan. Adjusted net losses were widening at RMB212 million, RMB322 million, RMB390 million and RMB393 million respectively.
On the positive side, while the losses have widened, the company's gross profit margin has been able to steadily increase. According to the prospectus, the gross profit margin of the company during the reporting period was 42.7%, 43.5%, 45.6% and 46.3% respectively, and the corresponding gross profit for the same period was 54.6 million yuan, 200 million yuan, 430 million yuan, 272 million yuan and 623 million yuan, respectively.
However, compared with the listed SenseTime, the gross profit margin level of the fourth paradigm is inferior. From 2018 to 2020, SenseTime's gross profit margin was 56.5%, 56.8% and 70.6%, respectively, and increased to 73% in the first six months of 2021.
In terms of R&D expenditure, in the first three quarters of 2018, 2019, 2020 and 2021, the fourth paradigm R&D expenditure was 193 million yuan, 416 million yuan, 566 million yuan and 844 million yuan, accounting for 151.2%, 90.6%, 60.0% and 62.8% of the revenue in the same period, respectively.
As of the end of the third quarter of 2021, the company's R & D staff has 1062 people, accounting for 72% of all employees. However, the Company expects that R&D investment will decrease as a share of total revenue in the future, due to the Company's expectation of greater bargaining power for its suppliers in outsourcing R&D services, and the fact that employee benefits and compensation will not increase in equal proportions as revenue increases.
The AI winter is not over yet
Driven by the wave of digitalization, artificial intelligence has rapidly penetrated into all walks of life, and gradually moved from theory to application and practice. But at the same time, burning money, losses, and listing difficulties have become lingering labels for AI companies.
In the domestic AI industry, in addition to the active layout of artificial intelligence by large factories such as "BAT", yuncong technology, Yitu technology, SenseTime technology, Megvii technology, as well as AI unicorns such as Cambrian, fourth paradigm and horizon have also grown up, known as the AI "Four Little Dragons".
However, these star companies in the primary market, the process of moving to the secondary market is very difficult. Cambrian, the "first stock of AI chips", was lucky, taking the lead in listing on July 20, 2020, and currently has a market value of 31 billion yuan and is still in a loss.
SenseTime also landed on the Hong Kong Stock Exchange on December 30, 2021 after experiencing a delayed listing storm, with a current market value of HK$218 billion. However, according to its previously released financial reports, from 2018 to the first half of 2021, SenseTime lost as much as 24.2 billion yuan.
Megvii is accepting the second round of inquiries from the CSRC, and Yitu Technology has also launched an impact on the Science and Technology Innovation Board, and then withdrew its listing application in June 2021.
Some senior investment bankers have analyzed that from the perspective of time nodes, AI companies are now in and will continue to be in a state of high investment and low income losses, which is not a good time to go public, and companies choose to queue up for listing, more out of helplessness.
Starting from 2021, the market has frequently reported serious news of layoffs, salary cuts, and overtime work of AI companies. There are rumors that the cloud was reported to be 30% off from the first half of 2021, which is another reduction after the last 30% layoff. Another insider said that the number of layoffs in the Yitu Medical Division exceeded 90%, which can be described as "surviving with a broken arm".
At the same time, some technical talents have left the AI industry and returned to academia. In July 2020, Ma Weiying, vice president and director of BYTEDance's AI Lab, left his post to join the Intelligent Industry Research Institute of Tsinghua University; in August 2021, Li Lei, director of ByteDance's AI Lab, left his post to join the University of California, Santa Barbara; in November, Qi Yuan, former vice president and chief data scientist of Ant Financial, joined Fudan University as the dean of Fudan Artificial Intelligence Innovation and Industry Research Institute.
Judging from the discussion in the industry, the investment and financing of the current market are concentrated in the basic layer and the technical layer, and the investment in the application level is obviously weak. In the future, AI companies need to continue to strengthen investment at the application level, reshape business models and promote the landing of AI scenarios, and ultimately achieve profitability.
In fact, looking at each AI start-up company, not only is it difficult to guarantee its own economic benefits, but also for customers who invest a lot of money, the return is also vague.
Taking the fourth paradigm as an example, an AI industry source who has had business dealings with large banks revealed that because the fourth paradigm provides the underlying technical capabilities, it is difficult for customers to use directly, and it is also necessary to spend extra money to find technology companies to help land.
This kind of low economic benefit and immature business model plagues most AI companies and hinders the entire industry from moving towards a healthy state of operation.
From the perspective of the development of the fourth paradigm alone, even if it successfully lands on the capital market, listing can only extend the "burning money" time, and it is still doubtful when it will be able to achieve profitability.
Note: This article is the original of Radar Finance (ID: leidacj). Unauthorized reproduction is prohibited.