Innovation Qizhi, who had submitted the form in June 2021, waited for half a year, the application materials on the Hong Kong Stock Exchange "expired" in December of that year, just after New Year's Day, before the 2021 annual financial report came out, Innovation Qizhi once again submitted the form to the Hong Kong Stock Exchange, but the listing process was exceptionally smooth. More than a week later, on 12 January, the HKEx website showed that Innovation Qizhi had passed the listing hearing. According to the announcement of the Hong Kong Stock Exchange on the 17th, the company plans to issue 44,744,400H shares at an offer price of HK$26.30-27.30 per share, of which 90% are international placements, and is expected to be listed on January 27.
Judging from the performance of the prospectus in recent days, as of January 18, Innovation Qizhi has recorded a total of 450 million yuan of margin subscriptions, and if it is calculated by the public offering part of the raised 122 million yuan, it has recorded a total of 2.7 times oversubscription, and the popularity of retail subscription has also exceeded market expectations.
Headquartered in Qingdao, the company, which is named AINNOVATION (AI+Innovation), mainly provides enterprises with full-stack AI products and solutions, has three independently developed proprietary AI platforms, serving system integrators and end customers in manufacturing, finance and other industries, while the digitization of manufacturing and financial industries is the focus of national industrial upgrading.
On January 12, the State Council issued the 14th Five-Year Plan for the Development of the Digital Economy, which clarified the guiding ideology, basic principles, development goals, key tasks and safeguard measures for promoting the healthy development of the digital economy during the 14th Five-Year Plan period. The "Planning" pointed out that during the "14th Five-Year Plan" period, the mainland's digital economy will shift to a new stage of deepening application, standardized development and inclusive sharing. Previously, in December, the Ministry of Industry and Information Technology and other eight departments jointly issued the "14th Five-Year Plan for Intelligent Manufacturing Development", which pointed out that by 2025, manufacturing enterprises above designated size will basically popularize digitalization, and key enterprises in key industries will initially achieve intelligent transformation.
The imminent listing of Innovative Qizhi also clearly shows the trend of Hong Kong stocks to accelerate the absorption of technology research and development projects after Eagle Pupil and SenseTime. In the past, the Hong Kong capital market still likes traditional industries, profitable companies, and the valuation multiple is not high, in recent years, the Chinese stocks listed in the United States have returned to Hong Kong for listing, setting off a wave of hong Kong stock new economy listing boom, during which also includes millet, Meituan and other companies directly to Hong Kong listing, but they still have a good level of profitability. The emergence of innovative Qizhi and SenseTime now directly pushes R&D "hard technology" companies to the big stage of Hong Kong's capital market, which is not only the transformation of these technology companies, but also an important signal for the change of the concept of Hong Kong's capital market now and in the future.
Anchoring the AI+ manufacturing industry is the key drop
The recent hit TV series "Win or Lose" not only reproduces the battlefield of IT companies in full swing in the past, but also indicates a new situation of enterprise-level market competition that is coming. With the deepening of the national digital economy layout into a broader industrial ecology, and the accelerated commercialization of artificial intelligence technology to meet the needs of all walks of life, the main force of the market is gradually shifting from consumer-facing (ToC) to enterprise-facing customers (ToB), and the enterprise-level digital market will once again face competition.
The artificial intelligence market in 2022, on the one hand, will further become an important driver of enterprise digital transformation. According to Frost & Sullivan, global IT spending (including spending on hardware, software, solutions, and industrial automation equipment) is expected to reach approximately $5,657.8 billion by 2025, and global AI-related spending is expected to reach $769.7 billion, accounting for 13.6% of total global IT spending. Among them, China, as a powerhouse in artificial intelligence, is expected to reach RMB 1,045.7 billion in 2025, accounting for 20.9% of the global artificial intelligence market.
Chinese Intelligent Market Size, Source: Innovative Qizhi Prospectus
On the other hand, enterprise-level artificial intelligence needs to be further integrated with traditional industries to have strong commercialization capabilities. At present, the participants in the industry are not only artificial intelligence technology-driven solution providers such as Innovation Qizhi, but also traditional IT companies and Internet companies, and the strength of enterprises in the face of competition in the industrial market shows their own tension.
Like the gradual end of hikvision, the first share of artificial intelligence security in the fields of public security, traffic police and other fields, such as xueliang project, and then with the perception of equipment to strongly cut into the field of digital transformation of smart cities and industrial enterprises, opening the company's second growth curve. The recent impact on the sci-tech board's holding subsidiary Fluorite Network shows that Hikvision has achieved a large market space for entering the smart home. SenseTime also has a major layout for computer vision applications in smart cities, smart life and other fields, in addition to its layout of automotive applications including autonomous driving. Another Hong Kong-listed artificial intelligence company, Eagle Pupil, focuses on the medical field, which is also one of the important areas of YITU.
Market share of China's AI technology-driven solution head provider in 2020, source: Innovation Qizhi Prospectus
Innovation Qizhi chose an industry that was rarely chosen by leading artificial intelligence companies at that time - manufacturing. We know that the main professional background of artificial intelligence scientists and engineers is IT-related disciplines, and they are more from cities, homes, and automobiles that can be accessed in daily life, and are less involved in the manufacturing fields of machinery, energy, and steel.
Kai-Fu Lee, chairman of Innovation Qizhi, said that China's huge manufacturing ecosystem has helped it become the "world's factory". Competition at home and abroad in the pursuit of production quantity and quality is fierce, and with the introduction of China's national "new infrastructure" policy, artificial intelligence will continue to play a key role in the process of China's economic reshaping and modernization. Traditional industries have formed an ecosystem of large-scale upstream and downstream synergy, and the technological revolution will lead to the improvement of the entire supply chain and even subvert the traditional model.
AI solutions in manufacturing are capable of undertaking tasks such as inspection, operation, monitoring, and transportation, which are typically solved by human beings. With AI solutions, the manufacturing workforce can be freed from repetitive and dangerous tasks and redistributed to perform jobs with higher economic outputs.
When obtaining the D round of financing, Xu Hui, CEO of Innovation Qizhi, also said: "China's manufacturing industry has a complete range of categories and a large industrial scale, and many fields are worth exploring and upgrading with artificial intelligence technology. Innovative Qizhi focuses on steel metallurgy, engineering construction, panel semiconductors, automotive equipment, energy and power, 3C high-tech and other sub-sectors, and launches a variety of industry-leading or industry-only artificial intelligence products and solutions, which have been recognized by many industry benchmark customers. ”
Today, in the manufacturing industry's artificial intelligence solutions, Innovation Qizhi has achieved 2.1%, which has gained a leading position in similar enterprises. The market development rate in this area, according to Frost & Sullivan's forecast, will grow at a compound annual growth rate of 48.3% from 2020 to 2025 and will reach 64.9 billion yuan in 2025, and even if the market share of Innovation Qizhi does not change by 2025, its revenue will be 7.13 times that of 2020.
Market share of AI solution providers in China's manufacturing sector in 2020, source: Innovation Qizhi Prospectus
The achievements of Innovation Qizhi in the manufacturing industry also complement its headquarters in Qingdao. As we all know, Qingdao has a relatively deep manufacturing foundation, and the Haier, Hisense, Tsingtao Beer, Double Star and AUCMA that we are familiar with are all produced in Qingdao. In recent years, in the face of the national deployment of digitalization, intelligent transformation and upgrading of the manufacturing industry, Qingdao has determined the main direction of attack with intelligent manufacturing, comprehensively promoted the upgrading of the city's manufacturing industry base and the modernization of the industrial chain, and accelerated the construction of the "World Industrial Internet Capital".
On March 29, 2019, Qingdao Jimo District, CICC Jiazi and Innovation Qizhi signed a strategic cooperation agreement to unite the three forces of "government, capital and enterprises" to jointly promote the landing of artificial intelligence in regional cities, empower local traditional industries to upgrade, and accelerate the pace of new and old kinetic energy conversion. According to the "New Qingji" report, one of the investors in the A, B and C rounds of innovation, CICC Jiazi, behind which is the CICC State Adjustment Science and Technology Innovation Fund, while the Jimo District Equity Investment Guidance Fund participated in the CICC State Adjustment Science and Technology Innovation Fund, with a subscribed capital of 200 million yuan.
The logic behind the three-year compound growth rate of 252.5%.
Many people know that Innovation Qizhi is a project directly incubated by the innovation workshop led by Kai-Fu Lee, Dr. Kai-Fu Lee has also served as the chairman of the company since its establishment in 2018, in the year before the establishment of Innovation Qizhi, Kai-Fu Lee has repeatedly said in public that there will be four waves of artificial intelligence, the first wave is Internet intelligence, Internet companies are rising because of user accumulation and mastery of big data; the second wave is business intelligence, based on the business process of business, activating all data to create value The third wave is the intelligence of the physical world, which captures various data through sensors and interconnects all things; the fourth wave is fully automatic intelligence, which will occur in the field of automation and autonomy.
Kai-Fu Lee once commented, "In the four waves of artificial intelligence development, the business of Innovative Qizhi spans the second wave of commercial intelligence and the third wave of physical world intelligence. Being able to contribute to users and the market with landable AI products, AI companies with commercial genes and engineering capabilities can better gain a foothold in these two waves. It is gratifying to see that Innovative Qizhi already has the above characteristics. ”
Backed by Kai-Fu Lee's strong R&D resources led by innovation workshop, the R&D strength of Innovative Qizhi can be imagined, in fact, the number of patents also proves this, at present, Innovation Qizhi has applied for 634 AI-related patents, successfully registered 126 AI-related patents, including 79 invention patents, and Innovative Qizhi has accumulated about 1,985 AI technology assets. However, R&D patents do not directly bring revenue, and what makes Innovative Qizhi transition from the laboratory to a commercial company, and what really lays the foundation for engineering and business is the AI platform.
The innovative AI platform enables it to accumulate a full set of AI technology assets, including datasets, operators, algorithm models, hardware components, design blueprints and software components, and continuously improve and upgrade the training model. At the same time, it also gives Innovation Qizhi the ability to customize various model combinations based on the differentiated needs of customers in various vertical fields in various industries, greatly improving the speed and efficiency of demand response. In fact, the AI platform of innovative Qizhi is not one, but three.
First, the Orion Distributed Machine Learning Platform, as an enterprise-level end-to-end machine learning platform, is used to train three algorithm models: computer vision algorithm models for defect detection, object localization, character or code recognition, and dimensional measurement; edge video analysis algorithm models for understanding video; and machine learning algorithm models for structured data for intelligent prediction and intelligent decision-making.
The algorithmic model can then be transferred to any of the three AI platforms for project delivery and field operation for its customers. Among them, the image processing algorithm model is delivered to the ManuVision machine vision intelligence platform, and the video processing algorithm model is delivered to the MatrixVision edge video intelligence platform. The former is a machine vision inspection software system designed to locate, measure, detect and identify common defects or key indicators, and provides a full range of solutions covering image annotation, deep learning model training, model testing, algorithm model pipeline and online inspection. The latter systematically combines edge computing and deep learning, and based on the innovative Qichi proprietary edge computing device, MatrixVision can perform tasks such as video stream decoding, image encoding and decoding, model conversion and migration, model deployment, and real-time inference.
The three AI platforms of Innovation Qizhi and their relationships, source: Innovative Qizhi Prospectus
Developing a rich portfolio of products and solutions from the AI platform and then outputting them into asset-based solutions (ABS) and rapid deployment products (RPPs), the role of these two modules is to be appropriately customized according to customer requirements and can be applied to various business scenarios.
Innovative Qizhi products and business association chart, source: Innovation Qizhi prospectus
For example, detecting defects on panel glass is a complex task, and the difficulty of the task changes as OLED panel glass manufacturing techniques change. Innovative Qizhi used the ManuVision platform to develop an OLED panel glass quality inspection solution through an AI algorithm model and deploy it on the customer's production line. Customers provide customers with a small number of sample images containing new defect patterns, and Innovative Qizhi improves the algorithm model based on a proprietary pre-trained AI model and a low-sample learning training algorithm.
Chuang chi chi has a leading position in the field of steel metallurgy in China in the field of steel metallurgy, with 40 of the world's top 50 steel mills. In MCC CCID Group, traditional molten iron transportation requires intensive labor, consumes a lot of human resource costs, and bears high safety risks. In 2020, Nova Qizhi significantly reduced the health and safety risks of field personnel through data reinforcement learning solutions on the Orion platform, with zero safety incidents that year. According to the feedback they received, one of the customers reduced labor costs by nearly 70% and saved more than 4 million yuan per year for a furnace site.
In this way, Innovation Qizhi first focused on reaching out to the top customers in the manufacturing and financial services industry, providing AI products and solutions for these lighthouse customers, and when they successfully set an example, on the one hand, they expanded to more scenarios of customers, and on the other hand, migrated the needs of the same customers, this model is called "1+N" cooperation method. Once this portfolio of products and solutions is established, Innovation Qizhi finds partners to form a network effect through the replicability of ABS and RDP, which is known as the "1*N" expansion method.
Following these two models, Innovation Chichi provided products and services to more than 400 customers, their manufacturing customers increased from 16 in 2018 to 93 in 2020, and the number of financial services customers increased from 2 in 2018 to 18 in 2020. According to public media reports, Chuangxin Qizhi has provided services for industry benchmarking enterprises including MCC CCID, Zongshen, CIMC, Baowu Iron and Steel, China Railway Fourth Bureau, China Textile Standard, Hon Hai Technology, Huadian Electric Power Research Institute, Mars, Postal Savings Bank, Pudong Development Bank, PICC Property & Casualty Insurance, Taikang and other industry benchmarking enterprises.
Innovative Chi Chi Intelligent Manufacturing System, for the digital transformation needs of manufacturing customers, source: Innovation Qi Zhi Prospectus
At this rate of customer growth, Innovation Qizhi's revenue increased from RMB37.2 million in 2018 to RMB229 million in 2019, and then to RMB462 million in 2020, with a COMPOUND ANNUAL GROWTH rate of 252.5% from 2018 to 2020. Revenue in the first three quarters of 2021 was 553 million yuan, exceeding the revenue of the whole year of 2020, of which the manufacturing industry was 281 million yuan. In terms of revenue, according to Frost & Sullivan, Chuangxin Qizhi is the largest AI technology-driven provider in China's manufacturing AI solutions market.
Compared with 2020, the main sources of revenue in the first three quarters of 2021 are automotive equipment and 3C high-tech, compared with the same period, the revenue of the predecessors in 2020 is only 981,000 yuan, and by 2021, it will jump to 116 million, and the latter will also go from 11.4 million to 77.85 million. The changes in the financial data of other manufacturing sectors served by Innovation Qizhi are not very obvious.
It should be noted that in addition to the increase in revenue, the gross profit margin of Innovation Qizhi decreased from 62.9% in 2018 to 31.3% in 2019 and 29.1% in 2020, of which the gross profit margin of products and solutions used in the OLED panel semiconductor industry was only 10.9% in the first nine months of last year. The main reason explained in the prospectus is that sales in 2018 were mainly based on software solutions, while AI products and solutions sold since 2019 have integrated hardware in addition to software. Since the higher cost of purchasing hardware and the proportion of relative revenue also jumped from 4.69% in 2018 to 46.98% in 2020, the more hardware components involved in the integrated solution, the often resulting in lower gross margins.
Innovative Qizhi quality inspection hardware, source: Innovative Qizhi website
We can't ignore that behind the innovative Chichi AI platform-based solutions, each comes with a series of hardware as a support. Before 2019, Innovation Qizhi was still based on software, and in 2019, Innovation Qizhi made strategic adjustments to face the needs of the manufacturing industry more broadly and deeply. The equipment of industrial enterprises is difficult to directly connect with software, that is, the last kilometer problem of artificial intelligence application landing, then a large number of hardware is added, so that the software is paired with compatible hardware, so that enterprises can use the innovative Chizhi software and hardware integrated artificial intelligence solutions out of the box, without having to do complex settings.
For example, at that time, the innovative Qizhi launched the ManuVision platform to be compatible with external hardware equipment such as industrial light sources, industrial cameras, manipulators, controllers, etc., to create an integrated automation solution for customers that integrates light (learning), machinery (machinery), electricity (gas), software (parts), and calculation (method), and can be flexibly integrated on the original production line of the enterprise, without special transformation of the production line, greatly reducing the difficulty and threshold of enterprise application of AI.
Integration of components into AI products and solutions, source: Innovation Chi chi prospectus
The hardware components of such a solution are usually customized according to customer requirements, and when the more customers put forward requirements, the more hardware will be matched, and the gross profit margin will become lower. This is not only the process of customer accumulation, but also the process of exploration, and they are involved in the manufacturing industry is not one, but more than five, they also need to run in with customers, continue to explore, merge similar items, the future may be further standardized, or reduce material loss.
Another sacrifice the team may have made was to voluntarily become a subcontractor in order to expand the market. We can see from the prospectus of Innovation Qizhi that in 2018, their end customers accounted for 84.7%, and system integrators accounted for only 15.3%, while by 2020, system integrators accounted for 76% and end customers accounted for 24%.
According to the definition of Innovation Qizhi, system integrators are mainly information technology service providers entrusted by end users to produce it information technology products or services/integration and management services. Innovation Wonder generally does not directly participate in the end-user bidding process of the system integrator, but acts as one of the subcontractors to meet the bidding requirements. Under this mechanism, Innovation Qizhi will purchase the corresponding software and hardware according to customer requirements, and then integrate them into the complete solution. Compared with the past of selling separate solutions to end customers, there is now an additional intermediate link, and the low gross profit margin is self-evident.
Innovation Chi Chi and Cognex released the industry's first industrial OCR intelligent handheld terminal based on deep learning algorithms, source: Innovation Qi Zhi website
The four-year loss of 1.1 billion yuan was mainly spent on the accumulation of talents, and it was necessary to continue to invest
In addition to the low gross profit margin, Innovation Qizhi has lost 1.1 billion yuan in four years, and the prospectus shows that the core of the main administrative and research and development expenses is employee compensation expenses and share-based payment expenses, which account for 72.3% in 2020. For some executives and R&D leaders, the latter is much higher than the former, for example, CEO Xu Hui was paid 1.619 million last year, while share-based payments were as high as 71.518 million. Excluding share-based payment expenses, financial costs of financial liabilities for redemption of shares, and listing expenses, the sum of the four years is only $430 million.
"Share-based payment expenses" are simply understood as share-based awards, and Innovation QiZhi clearly stated its intentions in the prospectus, "We believe that the granting of share-based compensation benefits is important to our ability to attract and retain key personnel and employees." It is understood that equity awards are more common among innovative companies, and the excessive expenditure of SenseTime, which also has a net loss, has caused controversy, and in the six months ended June 30, 2021, sensetime executives' "share-based compensation expenses" reached 510 million yuan, 374 million yuan and 304 million yuan, respectively.
Breakdown of components of Innovation Chi Chi General and Administrative Expenses, Source: Innovation Qi Zhi Prospectus
Behind the controversy, we must see that at present, in the field of artificial intelligence, especially in the field of intelligent manufacturing, the supply of talents is quite short, and excellent talents have a large bargaining space. According to the "2022 China Intelligent Manufacturing Technology Talent Insight" report jointly released by the Zhongguancun Industrial Research Institute and the LinkedIn, there is currently a great shortage of intelligent manufacturing technology talents in China, which may become a "card neck" problem that restricts development. Technical talents in this field often need to have cross-domain knowledge of digital technology and manufacturing, need to know how to work with machines or digital tools, and need to do a good job of "translation" between machines or digital languages and actual manufacturing scenarios.
At present, 52.8% of the R&D personnel of Innovation Qizhi are employed, and it plans to hire about 310 AI and software talents and industry experts in the next four years to strengthen the existing R&D projects and product levels, and recruit about 240 AI and software talents and industry experts to develop new features. We have seen that software engineers, algorithm engineers, and product managers have all received an average annual salary of more than HK$700,000, and researchers and scientists have reached an average of 900,000 and 1.2 million respectively.
Future Recruitment Plan for AI and Software Talents and Industry Experts (excerpt), Source: Innovative Qizhi Prospectus
Future Recruitment Plan for Cutting-edge AI Technology Research Talents (excerpt), Source: Innovative Qizhi Prospectus
Innovation Wonder said that at least 15 additional world-renowned AI researchers and scientists will be recruited to join the scientific research committee, a level of talent that, according to Frost & Sullivan, will have just over 4,000 worldwide by 2020.
It is worth noting that Innovation Qizhi will also strengthen its own sales capabilities with the goal of commercialization, especially with third-party system integrators and business partners, aiming to increase the penetration rate of industries and regions with strong demand for its products, and is expected to recruit a total of 400 sales staff in the next 4 years, with an average salary of between HK$500,000 and HK$1.4 million.
Future Recruitment Plan for Sales Talent (excerpt), Source: Innovative Chi Chi Prospectus
Compared with other entrepreneurial teams of scientists, Innovative Qizhi has a stronger commercial operation ability. This can be based on their previous resumes, holding senior leadership positions in various well-known high-tech and innovative enterprises, which can show strong business strategy ability, the ability to effectively understand customer needs, the experience of communicating with head customers, and the ability to integrate resources. The existing business results are a concentrated manifestation of these capabilities, and this competitive advantage may be strengthened in the next few years.
Source: Innovative Qizhi Prospectus, Data Ape Collation
When Kai-Fu Lee first handed over the form on the Hong Kong Stock Exchange in June last year, he wrote in the "Open Letter from the Management of Innovation Qizhi to Investors" that in China, we see a huge demand for artificial intelligence technology to drive industrial upgrading. We made a lot of progress in our early stages of technology development, and I understand very well the advantages and challenges of pure technology leadership in the process of technology commercialization. Later, Xu Hui joined, he has 22 years of experience in the B2B industry in China, working in Fortune 500 companies such as IBM, SAP, Microsoft and Wanda, and he became my co-founder and CEO, which allowed the company to find its helmsman.
"AI can't stay in the showmanship stage, can't stay in the lab, and must be deeply integrated with the industry in order to transform data value into commercial value." Xu Hui said this in an interview with a media outlet in 2019. We have seen that in the past two or three years, Innovative Qizhi has actually been practicing this route.
Kai-Fu Lee said that now that innovation Qizhi has realized our long-awaited dream, artificial intelligence, a revolutionary technology, has gone out of the academic circle, launching a new economic change of AI for industrial upgrading, which will change the world possible. The company continues to take "artificial intelligence empowers business value" as its mission, and is committed to helping Chinese enterprises transform in the upgrading of artificial intelligence and become the leader of AI technology.
Text: Luis / Data Ape