According to documents from the Hong Kong Stock Exchange, Yunzhisheng Intelligent Technology Co., Ltd. has recently submitted a listing application to the main board of the Hong Kong Stock Exchange. CICC and Haitong International are co-sponsors. As a domestic AI voice unicorn and Asia's leading AGI technology leader, this news has received close attention from the market.
Looking at Yunzhisheng's pre-listing shareholder lineup, it can be called a “gathering of stars”: including China Internet Investment Fund, JD, Qihoo 360, Qualcomm Venture Capital, Qiming Venture Capital and other domestic and foreign first-tier war investment institutions and national team funds. Up to now, Yunzhisheng has completed the D3 round of financing, raising more than 700 million yuan.
With the rapid development of artificial intelligence technology and the explosive growth of application demand in various industry scenarios, intelligent voice is becoming increasingly heated competition on the racetrack as a key interface for human-computer interaction in the intelligent era. Regardless of whether you look at the potential size of the market or the company's steady fundamentals and deep technology accumulation, Yun Zhisheng has the potential for high popularity and high growth after officially entering the Hong Kong stock market.
Overall improvement of financial indicators lays the foundation for sustainable growth
According to public information, Yunzhisheng is an artificial intelligence enterprise with independent intellectual property rights. The company's history can be traced back to June 2012. Currently, it is mainly based on the Yunzhi brain technology platform, focusing on deepening the two key application scenarios of smart living and smart medical care, and providing customers with highly competitive products and solutions.
For a company that has entered the open market, performance is unquestionably the best litmus test and the strongest proof that the company has shown fundamental resilience. Meanwhile, in terms of revenue and profit indicators, Yun Zhisheng has already handed over a “report card” with excellent performance.
During the reporting period, the company's revenue grew from 261 million yuan in 2020 (unit is RMB, same below) to 600 million yuan in 2022, with a compound annual growth rate of 51.7%; gross profit increased from 89.25 million yuan in 2020 to 240 million yuan in 2022, with a compound annual growth rate of 63.9%, both recording rapid growth. Meanwhile, gross margin, which is one of the important profit indicators, has steadily increased. Gross margins in 2020, 2021, and 2022 were 34.2%, 31.8%, and 39.9%, respectively.
If you broaden your horizons from past financial indicators to the longer term future, then you'll find thatBehind the rapid growth of Cloud Intelligence's performance are deep technology accumulation, stable customer base, and efficient cost advantages, which have jointly set the tone for sustainable growth.
In terms of artificial intelligence technology, cloud intelligence can be called “early birds fly first.” According to the prospectus, the company began launching a voice cloud platform as early as 2012, and was one of the first companies to apply deep learning algorithms to commercial speech recognition; built a “cloud core” technology architecture scheme in 2014 to complete the commercialization of intelligent voice solutions; began building an Atlas artificial intelligence infrastructure based on intelligent computing clusters in 2016 to efficiently and dynamically schedule powerful computing power for machine learning tasks; and launched UniCore, a large language model with hundreds of millions of parameters in 2019.
According to the Zhitong Finance App, by 2022, the company had sold about 13 million consumer-grade AI chips and commercialized and put automotive chips into use; in 2023, the company also launched its own big language model Shanhai Big Model with 60 billion parameters, and applied the successful delivery of smart living solutions to subway lines in cities such as Guangzhou, Shenzhen and Qingdao.
Positive technological progress is inseparable from continuous investment in research and development. In 2020, 2021 and 2022, Cloud Zhisheng's R&D expenses were 189 million yuan, 188 million yuan and 287 million yuan respectively, accounting for 72.6%, 41.3% and 47.8% of the company's annual revenue respectively. As of December 31, 2022, the company's R&D team had 281 people, accounting for more than 70% of the total number of employees, and the company had 288 registered patents, 563 patent applications, 235 trademarks and 195 software copyrights in China.
At the business level, stable and high-quality customer resources not only show the market's recognition of the company's strength, but also the foundation for the sustainable growth of the company's performance in the future.
At present, Cloud Zhisheng has reached strategic cooperation with telecom and other operators, and downstream customers are involved in multiple vertical industries. The customer groups include Gree for home appliances, Ping An Technology for health care, Peking Union Medical College Hospital for medical services, and China People's Insurance for health insurance. From 2020 to 2022, revenue from the top five customers accounted for 40.1%, 34.4% and 30.8% respectively. The customer structure is diversifying, which also shows that the company's bargaining power is steadily improving.
In terms of cost, benefiting from the Atlas artificial intelligence infrastructure, cloud intelligence brain technology platform, and highly standardized products and solutions, combined with the scale effect of business growth, the company has achieved remarkable results in reducing costs and increasing efficiency in recent years.
In the period 2020-2022, Cloud Intelligence's sales and marketing expenses were 64.2 million yuan, 60.5 million yuan, and 46.1 million yuan respectively, showing a gradual downward trend. Meanwhile, with the expansion of business scale and continuous increase in brand awareness, the number of customers increased from 459 in 2020 to 538 in 2022.
Behind many signs of optimism, with technology as the core driving force, Cloud Intelligence's artificial intelligence business has already shown full growth momentum. It is worth mentioning that at a time when AI big language models are rapidly rising, Cloud Intelligence also has many bright spots in terms of big language models and their commercialization path.
Deeply cultivating key industry scenarios The potential of the Shanhai model needs to be unleashed
Looking forward to the future of the industry, the artificial intelligence circuit is ushered in new opportunities, and the potential market space is broad. According to data from Frost & Sullivan, driven by economic growth and changing customer needs, China's artificial intelligence solutions market increased from 42.2 billion yuan in 2019 to 93.7 billion yuan in 2022, with a compound annual growth rate of 30.5%. It is expected to reach 633.2 billion yuan in 2028, with a compound annual growth rate of 37.5% from 2022 to 2028.
In segments such as AI, smart living, and smart healthcare, the cloud intelligence industry has a strong dominant position. According to information from the Zhitong Finance App, according to Frost & Sullivan's data, in terms of revenue, Yunzhisheng was the fourth largest AI solution provider in China in 2022, and grew the fastest among enterprises with annual revenue exceeding 500 million yuan; in the same year, in terms of revenue, the company ranked third in the smart living and smart hospital industry in China.
Specifically, according to Frost & Sullivan's data, according to revenue statistics for 2022, the company's market share as a smart lifestyle artificial intelligence solution provider is 8.4%; at the same time, the company's market share as a smart hospital AI solution provider is 6.6%. Both are at the forefront of the industry and have a steady leading edge.
Smart living solutions cover a wide range of application scenarios such as residences, commercial spaces, hotels and transportation. They aim to bring convenience to people's daily lives, improve the quality of life of consumers, and create a smoother user experience. Benefiting from multiple favorable factors such as the continuous deployment of 5G networks and the development of edge computing technology, the penetration rate of smart IoT will gradually increase in 2023, and shift from consumer IoT limited to households to a wider range of industrial IoT, smart buildings, and smart cities.
Smart healthcare is an important application in the field of people's livelihood. With the introduction of many digital health care policies and the deepening of new medical reforms, the transformation of the healthcare industry to digital informatization has become the general trend. However, smart healthcare in China is still in its infancy. Given the general deficiencies of the industry, such as insufficient number of medical professionals and low degree of digitalization, the use of advanced technologies such as AI and big data to empower the medical industry will have a broad market space.
Whether it's the smart lifestyle or smart medical industry, these segments have the characteristics of low concentration, many participants, and scattered competitors. Based on the Cloud Intelligence Brain Technology Platform, the company provides highly competitive products and solutions for a wide range of application scenarios in smart living and smart healthcare.
Cloud Intelligence products and solutions mainly include the two major sectors of smart living and smart healthcare. In the existing product portfolio, the company has various delivery and deployment methods such as integrated chips, edge SDKs, public and private cloud deployments, and provides developers and customers with professional capabilities based on interactive AI. As of December 31, 2022, the company's solutions have been used in a wide range of vertical industries, including healthcare, home, commercial space, hospitality and transportation.
An intelligent computing cluster with over 200 PFLOPS of computational power is a key cornerstone of the company's AI models and solutions. It has more than 1.6 PB of storage capacity, can efficiently and dynamically schedule thousands of GPUs for parallel computation and seamless dynamic expansion; for large language models, the intelligent computing cluster optimizes the parallel mechanism in depth engineering efficiency, combining advanced AI algorithms such as convolutional neural networks (CNN), generation of adversarial networks (GAN), and reinforcement learning (RL). In 2018, the company won the top two places in the WMT 2018 World Machine Translation Technology Evaluation, and launched the big language model Shanhai Big Model in 2023.
As the core of the company's key technology platform “Cloud Intelligence Brain”, the “Big Mountain and Sea Model” has gone through extensive text and code pre-training and has more than 60 billion parameters.It includes general capabilities such as language generation, semantic understanding, knowledge questions and answers, logical reasoning, mathematical calculation, code generation, and security compliance. On the basis of general capabilities, the Shanhai Model can optimize industry implementation capabilities through plug-in expansion, industry enhancement, and customization.
For example, for specific industries such as the medical field, through incremental training such as high-quality medical textbooks, clinical guidelines, medical records, etc., the Shanhai University Model passed the US Physician License Examination and obtained 81 points in the MedQA data set, surpassing GPT-4, and the knowledge and ability to solve professional problems improved markedly.
In terms of versatility,The Shanhai model is based on a framework of large-scale generative pre-training and command tuning. It has strong performance in natural language, general knowledge, reasoning and rapid learning. It is suitable for handling complex industry application scenarios and improving commercialization efficiency;
In terms of adaptability,Through incremental training of industry data, command fine-tuning, related plug-ins and AI components, it is possible to enhance the application of the Shanhai model in industry-specific scenarios without retraining the model, so that the Shanhai model integrates with the customer's own knowledge base, real-time information and ready-made APIs, and can adapt to highly specialized scenarios;
In terms of efficiency,By quickly performing task data adjustment procedures under a unified adaptation framework, it is possible to greatly reduce the human investment required during model retraining and process redevelopment, handle the mainstream needs of various industries, make AI solutions inherently highly adaptable, and greatly improve R&D efficiency.
According to information from the Zhitong Finance App, the Cloud Zhisheng Mountain and Sea model has now gone deep into specific scenarios such as smart medicine, knowledge management, quality control, and smart marketing to explore more possibilities for deep integration with specific scenarios. With the accelerated iterative evolution of the Shanhai model and the continuous enrichment and expansion of scenario applications based on the Shanhai model, it is expected to contribute new increments to the company's performance.
Summarize
Jade is looking for a good price in the mist, and flies when it is in the mist. As of 2023, the artificial intelligence circuit is still experiencing the “cold and cold alternation” of a boom in capital speculation and application implementation difficulties. There are also opportunities in challenges and changes. In an environment full of uncertainty, with a sound business and technology foundation, Cloud Intelligence is expected to achieve overall quantitative and qualitative growth, accelerate performance, and bring more imagination to the industry.