Recently, according to the latest news from the Shanghai Stock Exchange, Yuncong Technology disclosed the third round of inquiries and responses to the public. At the same time, the company was quickly arranged to attend the meeting and will usher in the IPO exam on July 20.
Yuncong Technology, Yitu Technology, SenseTime Technology, and Megvii Technology are known as the “AI Four Dragons”. In May, Megvii Technology announced the abandonment of the Hong Kong stock IPO, and then on July 2, Yitu Technology, which sprinted for the IPO of the Sci-tech Innovation Board, was changed to a terminated state, and SenseTime began to “be silent” about its listing. This casts a shadow on the cloud from the technology market.
An investor who has been tracking artificial intelligence for a long time pointed out to the 21st Century Business Herald that the peak period of AI industry financing is mainly concentrated before 2018. In 2018, the AI industry completed a total of 523 financings with a total amount of approximately 66.71 billion Yuan. By 2020, this set of numbers will drop to 305, 24.33 billion Yuan. “The artificial intelligence industry is difficult to research and develop technology and high investment in research and development. It requires companies to continuously invest in capital costs. Short-term output returns and profits are unrealistic. However, the capital market pays attention to profit first. AI companies have been losing money for many years and forced to land capital. The market may cause a bloody listing.”
According to industry statistics, nearly 90% of domestic artificial intelligence companies lost money in 2018, and the other 10% barely maintained a breakeven. Among the “AI Four Dragons”, Yuncong Technology has a net loss of nearly 2.3 billion Yuan from 2017 to the first half of 2020, and Yitu Technology has a total loss of nearly 7.3 billion Yuan from 2017 to the first half of 2020. Megvii Technology has a net loss of nearly 7.3 billion Yuan from 2017 to the first half of 2020. In the first three quarters of 2020, the total loss has exceeded 13 billion Yuan.
In China, artificial intelligence is penetrating into various vertical fields. Among them, the medical and health field accounts for up to 22%. The above-mentioned AI companies are also involved. As early as the second half of 2016, Yitu Technology has announced its entry into the medical market, but In terms of expanding medical resources, it is struggling. Yitu Medical, a subsidiary of Yitu Technology, has now cancelled its medical branches in Wuhan, Xi’an, Chengdu and other places, and related medical procurement and sales departments have been disbanded.
Well-known foreign technology companies such as Google, IBM, Apple, etc. are also deploying in the field of medical AI, but many innovative businesses are also at a loss.
The above-mentioned investors pointed out that the future development prospects of AI cannot be denied, but this is a long-term process, and it also requires the ability to integrate with various application fields and seek effective business models. “During this period, there will be a stage of squeezing bubbles. Whoever can find a suitable application scenario and business model will eventually have the possibility of surviving.”
In November last year, Yitu Technology took the lead in submitting an IPO application on the Sci-tech Innovation Board, which impacted China’s “first AI share”. As one of the first “AI Four Dragons” to impact a shares, Yitu Technology’s IPO process has attracted market attention.
Although they are all artificial intelligence companies, the business positioning of the “AI Four Little Dragons” is different.
Yitu Technology locates AI chips, with artificial intelligence chip technology and algorithm technology as the core, develops and sells artificial intelligence solutions including artificial intelligence computing hardware and software.
Cloud is positioning human-machine collaboration from science and technology, and has added new investment in robotics and Internet of Things technology research in an attempt to build more standardized AI products.
While Megvii Technology cuts into AIoT and regards logistics business as a future growth point. Its main business includes personal IoT solutions, urban IoT solutions, and supply chain IoT solutions.
SenseTime, considered the leader of the “AI Four Little Dragons”, positioned its AI factory and proposed a platform strategy of “1+1+X”, where 1 represents R&D and technology industrialization, and X represents empowerment for all industries.
According to different application scenarios, Yitu Technology’s main business is divided into two categories: smart public services and smart commerce. It currently serves more than 800 governments and enterprises in more than 30 domestic provinces, autonomous regions, municipalities directly under the central government, and more than 10 foreign countries and regions. End customers provide products and solutions.
According to Yitu Technology’s prospectus, its revenues for 2017, 2018, 2019 and the first half of 2020 were 68.718 million Yuan, 304 million Yuan, 717 million Yuan and 381 million Yuan, respectively. During the same period, Yitu Technology’s net profits attributable to its parent were 1.166 billion Yuan, 1.161 billion Yuan, 3.642 billion Yuan and 1.299 billion Yuan. As of the end of June 2020, the accumulated unrecovered losses exceeded 7.2 billion Yuan.
This time, Yitu Technology took the initiative to withdraw the order, which may be related to its three-and-a-half-year loss of 7.2 billion Yuan.
After a loss of 7.2 billion in three and a half years, why did Yitu Technology burn so much money? The 21st Century Business Herald reporter found three main reasons after combing. Firstly, because it is a technology-driven enterprise, in order to seize industry development opportunities, a large amount of resources were invested in R&D and innovation, market development, R&D, and sales expenses during the reporting period were 1.56 billion Yuan, 574 million Yuan, 1.075 billion Yuan, and 540 million Yuan.
In terms of R&D, Yitu Technology’s R&D expenses during the aforementioned reporting period were 101 million Yuan, 291 million Yuan, 557 million Yuan, and 381 million Yuan, accounting for 146.94%, 95.77%, 91.69%, and 100.10% of each period’s operating income, respectively. As of June 30, 2020, there are 837 R&D personnel, accounting for 55.54% of the total number of employees.
Although in terms of research and development results, Yitu Technology has used its artificial intelligence chips to “search” and become one of the few companies that have achieved product tape out and large-scale applications among artificial intelligence chip startups. However, according to Yitu Technology’s reply to the Shanghai Stock Exchange’s inquiry letter, the “search” chip was not originally developed by Yitu Technology itself, but from Yizhi Electronics, an equity-related company.
The prospectus disclosed that in April 2020, Yitu acquired a 22.38% stake in Yizhi Electronics from Yunfeng Xincheng, Sequoia Jiasheng, Hillhouse Zhicheng, and Hillhouse Yuying. Inquiry of industrial and commercial information shows that Yitu currently holds 28.92% equity of Yizhi Electronics and is the second largest shareholder of Yizhi Electronics, second only to its founder Xu Ruhao. According to the WEMONEY Research Office, the transaction caused Yitu’s goodwill to increase sharply from 0 to 1.496 billion. At the end of June 2020, Yitu’s total assets were 4.798 billion, of which goodwill accounted for 31.19%, and the risk of impairment was greater.
Although a lot of resources have been invested in R&D and innovation, Yitu Technology’s R&D investment still lags far behind domestic and foreign industry giants. In recent years, international giants such as Google and NVIDIA have deployed artificial intelligence technology research and development. Artificial intelligence computing power products and systems are iteratively fast and require continuous investment in innovation. However, Yitu Technology’s financing channels are relatively simple, and its financial strength and R&D investment are far behind the industry giants.
In addition, there is also a significant gap in business scale. Yitu Technology has been established for a short time, its sales network has not yet been fully developed, and its business scale is relatively small. In contrast, international giants such as Google and NVIDIA have more mature sales networks, product sales scale and market visibility.
Secondly, due to the increase in the overall valuation of the AI industry, the corresponding increase in the fair value of preferred stocks has led to a book loss, resulting in losses in fair value changes of 983 million Yuan, 545 million Yuan, 2.619 billion Yuan and 936 million Yuan in each period. Prior to this, Yitu Technology has achieved a valuation of more than 10 billion Yuan through 15 financings. As its valuation continues to rise, the increase in the fair value of preferred stocks has also been reflected in accounting liabilities. As of the end of June 2020, Yitu Technology’s asset-liability ratio reached 252.28%.
In this sci-tech innovation board listing, Yitu Technology’s new shares accounted for no more than 15%. It is expected to raise 7.5 billion, and the estimated valuation is about 50 billion. Although Yitu Technology promises not to reduce its holdings for 3 years after listing, it will not reduce its holdings if it is not profitable. However, judging from the status quo of multiple “suspension” to the final “termination”, the supervision does not agree with its valuation and its practices.
In addition, from the source point of view, Yitu Technology’s lack of profitability is the third largest cause of losses. During the reporting period, although Yitu Technology’s gross profit margin showed an overall upward trend, Yitu Technology’s gross profit margin was lower than the industry average. In 2017, 2018, and 2019, Yitu Technology’s gross profit margin was 57.39%, 54.55%, and 63.89%, while the average gross profit margin of the same industry was 86.44%, 85.31% and 73.45%, respectively. Yitu Technology’s gross profit margin continued to be significantly lower Average gross profit margin in the same industry; although Yitu Technology’s gross profit margin from January to June 2020 is slightly higher than the average gross profit margin of the same industry, there is still a big gap compared to Hongsoft Technology and Cambrian.
In recent years, Yitu Technology’s accounts receivable has continued to grow, and the risk of bad debts has increased, which may have an impact on its cash flow. According to the Institute of Listed Companies in the Financial Industry, from 2017 to the first half of 2020, Yitu Technology’s accounts receivable and bills receivable totaled approximately RMB 30 million, RMB 253 million, RMB 567 million, and RMB 695 million, respectively accounting for revenue 43.48%, 83.22%, 79.08%, 182.41%, showing a clear upward trend. However, the account receivable turnover rate continued to decline, from 2.68 times/year in 2017 to 0.57 times/year in the first half of 2020.
Although it is at a loss, the valuation of Yitu Technology is not low. According to the “2020 Hurun Global Unicorn List”, Yitu Technology is valued at 14 billion yuan. While continuing to lose money, superimposed on high R&D investment and pressured cash flow, whether Yitu Technology will continue to be favored by investors is full of uncertainty.
On July 15, a reporter from 21st Century Business Herald also called Yitu Technology to ask whether it would restart the IPO in the future, but the other party did not give an exact reply.
Finding an effective business model is critical
In fact, the entire AI industry continues to lose money. At the end of 2020, An Hui, deputy chief engineer of the CCID Research Institute of the Ministry of Industry and Information Technology and secretary-general of the Artificial Intelligence Industry Innovation Alliance, once stated that nearly 90% of artificial intelligence companies in the world are still at a loss, and more than 90% of the companies in the AI industry chain are also at a loss .
Judging from the prospectus, AI companies get together to go public, mainly to ease financial pressure. It is understood that Yitu Technology lost more than 7 billion Yuan in three and a half years, Megvii Technology lost more than 13 billion Yuan in four years, and Yuncong Technology also lost more than 2 billion Yuan in three years. With large-scale losses, AI companies need more funds to “make up blood”, and listing and financing is an important way. Therefore, other “AI three little dragons” SenseTime Technology, Megvii Technology, and Yuncong Technology are also actively accelerating the pace of listing.
But the road to an IPO is not simple. Lei Megvii Technology revealed that it aimed to go public in 2017, and the IPO in Hong Kong was postponed in 2019, and then in May of this year, it officially announced the termination of its Hong Kong stock listing plan. Megvii Technology has been walking this difficult road for 3 years.
Yitu Technology took the initiative to withdraw the order after nearly 8 months of preparations, and its IPO on the Science and Technology Innovation Board also failed. In order to reduce costs, Yitu Technology adopted measures related to salary cuts and substantial layoffs in the first half of this year. Now the total number of employees has been reduced to about 500. According to people familiar with the matter, the medical business department that Yitu Technology used to focus on has become the “hardest-hit area” for this layoff, with a reduction of about 70%.
It is reported that in 2018, 2019 and the first half of 2020, YITU’s medical and health application scenarios achieved revenues of 99,100 Yuan, 5.5973 million Yuan and 5.6267 million Yuan respectively. According to Yitu Technology, since artificial intelligence technology has not yet achieved large-scale commercial use in the medical field, the revenue during the reporting period was not large, accounting for a small proportion of operating revenue.
In addition, Yuncong Technology was also reported to have laid off 30% of its staff. This is another layoff after the last 30% layoff. From 2017 to 2019, the annual operating income of Yuncong Technology was 64 million Yuan, 484 million Yuan, and 807 million Yuan. In terms of profitability, the cumulative loss in the last three years exceeded 2 billion Yuan, but it was compared with other “AI three little dragons”. In comparison, Yuncong Technology’s total loss is not outstanding.
Therefore, finding an effective business model is crucial.
“When the entire industry suffers huge losses, especially when companies that have become the ‘AI Four Dragons’ are dying, we really need to calm down and think. Can companies survive only by AI services?” July 10, a domestic company senior executive of the listed companies of Intelligent Networking commented in the personal circle of friends.
As an underlying technology, AI should have the ability to integrate with various application fields, which will help it seek effective business models.
But when AI companies go deep into specific industries, they face greater challenges. “Artificial intelligence is currently mainly a new system based on deep learning algorithms. It is a direction developed in IT technology. When it is implemented, the compatibility and acceptance of existing software and hardware systems must be considered. Commercial advancement is not imagined. In an interview with the media, a programmer who has studied artificial intelligence technology for a long time said that the current AI technology is difficult to cope with the complexity of business and human nature. Therefore, when AI artificial intelligence companies enter the depths of the industry, they often It is difficult to obtain bargaining power and pricing power.
According to a research report by iiMedia Consulting, from the perspective of scenario implementation, AI technology is currently only implemented in areas with a foundation of digitalization and standardization in security, finance, etc. The commercialization effect in other areas is not yet satisfactory.
Where is the medical AI company going?
Among various vertical industries, artificial intelligence penetrates more in the fields of healthcare, finance, commerce, education, and security. Among them, the medical and health fields account for up to 22%, followed by the financial and intelligent commercialization fields at 14% and 11%.
According to IDC data forecasts, the total value of the artificial intelligence application market will reach 127 billion US dollars by 2025, of which the medical industry will account for one-fifth of the market.
With the continuous expansion of the market size of the medical AI field, its market entrants are also increasing. Well-known foreign technology companies such as Google, IBM, Apple, etc. are also deploying in the field of medical AI. But it is more difficult to get a share of the medical AI field. In November 2018, Google established “Google Health“, which merged DeepMind Health, the health department of DeepMind, and the team responsible for advancing the “Streams” medical APP. However, according to the 2021Q1 quarterly financial report, its innovative businesses such as artificial intelligence DeepMind and intelligent medical Verily are still at a loss.
At the business level, Google’s key health product “Diabetic Retinopathy Screening” has become “unacceptable” in practical applications. Google once said that the AI algorithm can make this tool accurate to 90%, and theoretically results can be obtained in a few seconds, “enough to be comparable to the diagnosis results of ophthalmologists.” However, in practical applications, due to network problems, it takes longer and the accuracy rate is not as expected, so it is difficult to land.
In addition, the effectiveness of IBM’s deployment of the medical AI track is not satisfactory. IBM has spent more than 4 billion US dollars to acquire Waston and set up Watson Health. But its annual revenue is only 1 billion U.S. dollars, and it has yet to make a profit. In addition, the industry continues to circulate news that IBM intends to sell Watson Health.
In fact, in the current medical AI field, companies that have truly landed and successfully listed are basically in the two directions of “big data management” and “voice entry.” However, clinical data involves patient privacy and is difficult to communicate and share in hospitals. This is the biggest obstacle facing the development of AI medical care.
In addition, standardized data specifications are also an important issue. Medical AI faces some unique and difficult obstacles in the world: the sensitivity of medical data and strict privacy protection regulations limit the collection of high-quality aggregated data required by AI medical treatment.
Our country’s medical AI companies mainly lay out the field of image recognition. According to industry statistics, there will be a total of 129 artificial intelligence medical companies in China in 2020 (excluding companies that focus on genetic testing technology). Among them, the number of companies in the medical imaging field is the largest, reaching 55, accounting for 42.6% of the total number of artificial intelligence medical companies. From the perspective of the establishment, no new companies with medical artificial intelligence as the main business will be established in 2019 and 2020. Most companies are working hard to find the possibility of their own landing and large-scale commercialization in various scenarios.
However, it is undeniable that medical AI has become an important direction for the development of medical technology in the future. The McKinsey Global Institute predicts that the large-scale use of artificial intelligence to diagnose diseases may not happen too quickly. Even the giants that have already entered the game are only entry-level. This does not prevent AI from successfully infiltrating and becoming the underlying technology of medical care. Like the previous IT technology.
The above-mentioned investors pointed out that as an underlying technology, AI’s ability to integrate with various application fields will help it to seek effective business models. For example, when entering the medical and health field, artificial intelligence itself also has strong medical attributes, which also means the initial investment is large and the return period is long. “AI is the direction of future development without doubt, but it still needs a long-term development process. During this period, there will be a stage of squeezing bubbles. Who can find suitable application scenarios and business models will eventually have the possibility of surviving.”