laitimes

Cloud from technology staged a "red start", Zhou Xi is still mired in losses

Produced by | Bullet Money

Author | Yin Taibai

Edit the | Egg total

American editor | Qianqian

On January 30, coinciding with the first day of the A-share market opening after the Spring Festival, a news from overseas made a listed company stage a "good start".

Recently, Microsoft announced that it will invest billions of dollars in OpenAI, which it has invested in twice in 2019 and 2021. ChatGPT, a chatbot developed by OpenAI, exploded at the end of 2022, making many investors and institutions more optimistic about the landing and commercialization prospects of AI technology.

This news spread quickly in the industry, and also made the stock price of A-share artificial intelligence company Yuncong Technology (hereinafter referred to as "Yuncong") rise on the first day of the market, soaring 19.99% in a single day to close at 20.23 yuan / share.

According to the observation of "Bullet Finance", on January 3 this year, Yuncong's stock price was 15.99 yuan / share, and as of the close of trading on January 31, its stock price was 21.02 yuan / share, and the stock price rose by 31.5% in one month.

This trend of sharp rise at the beginning of the year has attracted everyone's attention, after all, investors were not satisfied with Yuncong's performance in the secondary market. Before the Spring Festival, one investor complained in a post on the stock exchange platform: "Yuncong has fallen into 'cloud technology'. Some investors also left messages that "have left the market with pain" and "never again".

Now, with Microsoft's additional investment in OpenAI, Yuncong has ushered in a "spring blossom" in the secondary market.

On the investor interactive platform, someone asked Yuncong: "Recently, Microsoft invested an additional 10 billion yuan in OpenAI, and BuzzFeed soared by more than 100% due to the use of ChatGPT, do you have technical reserves or product planning in terms of the Chinese version of ChatGPT or large models?" ”

In response, Yuncong responded: "With the release of OpenAI super dialogue model ChatGPT, the industry has widely felt the huge opportunities brought by pre-trained large models to the AI industry... The company's philosophy is largely consistent with OpenAI. ”

Obviously, against OpenAI, there will be huge business opportunities in the cloud. However, it still faces a major challenge that needs to be solved urgently, that is, the company's huge losses.

Yuncong released its 2022 performance forecast on January 30, announcing that it is expected to have a net loss of 785 million yuan to 934 million yuan in 2022, and the loss will expand by 24.18% to 47.75% compared with the same period of the previous year.

Obviously, Yuncong's founder and chairman Zhou Xi has not been able to prevent Yuncong's losses from continuing to expand, and its listing prospectus has written that it expects to be profitable in 2025. So, with only two years remaining, how likely is it that the cloud will turn a profit?

Can the "human-machine collaborative operating system and the first curve and second curve" two types of solutions" mentioned by Zhou Xi in his new vision in 2023 and the concept of "meta-universe" of the fire really lead the cloud out of the quagmire of continuous losses?

1. Huge losses continue

According to a record of investor relations activities disclosed by Yuncong, half a year after Yuncong's listing, the most frequently asked questions by investors are still when to make a profit, how the technology is implemented and how to solve the phenomenon of declining performance.

Even if it has successfully knocked on the door of the capital market, many of Yuncong's hard injuries have not been effectively improved.

According to the third quarter of 2022 financial report of Cloud, its total revenue not only fell sharply, but also the net loss attributable to the parent continued to expand.

Specifically, the total revenue of cloud from the first three quarters was 456 million yuan, a year-on-year decrease of 41.53%; The net loss attributable to the parent was 589 million yuan, an increase of 42.62% over the same period last year.

Compared with the first three quarters, the decline in total revenue and the expansion of net loss attributable to the parent in the third quarter of Yuncong were more obvious, and in the third quarter of 2022, the total revenue of Yuncong was 64.7903 million yuan, a year-on-year decrease of 80.07%; The net loss attributable to the parent was 264 million yuan, a year-on-year increase of 457.02%.

In this regard, Yuncong explained in the financial report that the decrease in total revenue was mainly due to the impact on market demand and huge uncertainty caused by the repeated domestic new crown pneumonia epidemic and economic slowdown, and the delivery and acceptance progress of some projects were delayed.

Photogram / Photogram, based on VRF protocol

The net profit loss attributable to the parent was mainly due to the year-on-year decline in operating income, the year-on-year increase in R&D investment and the year-on-year decline in government subsidies.

In short, after the listing, the cloud still does not have hematopoietic ability, and in the case that open source cannot be opened and throttling, the cloud can only sink deeper and deeper into the quagmire of losses.

In fact, huge losses have become the "sword of Damocles" hanging over the heads of many AI companies such as Yuncong.

According to the prospectus, from 2019 to 2021, Yuncong's net loss attributable to the parent was 640 million yuan, 813 million yuan and 632 million yuan, respectively, and if you add the net loss attributable to the parent of 589 million yuan in the first three quarters of 2022, it has a total loss of 2.674 billion yuan in the past four years.

However, compared with other members of the AI "Four Little Tigers", Yuncong's loss is the lightest.

From 2017 to 2019 and the first nine months of 2020, Megvii had a net loss of RMB13.065 billion, YITU had a net loss of RMB7.268 billion from 2017 to the first half of 2020, and SenseTime had a net loss of RMB37.384 billion from 2019 to the first half of 2022.

Knowing that the huge losses are unsustainable, Yuncong predicted the operating conditions in the next five years (2021-2025) in the prospectus, according to the prediction results, the total revenue of Yuncong will reach 1.201 billion yuan, 1.910 billion yuan, 2.542 billion yuan, 3.259 billion yuan and 4.064 billion yuan respectively, with a compound growth rate of 35.64%, and the time node to achieve a turnaround is 2025.

However, combined with the performance and trends of cloud since its listing, it will not be easy to turn losses into profits in 2025.

On the one hand, from 2019 to 2021, the total revenue of Yuncong was 807 million, 755 million and 1.076 billion yuan, respectively, with a compound growth rate of only 15.47%, which is far from the target of 35.64%.

On the other hand, from 2019 to 2021, the comprehensive gross profit margin of Yuncong was 40.89%, 43.46% and 37.01%, respectively, and there was a downward trend.

In comparison, SenseTime's comprehensive gross profit margins for the same period were 56.8%, 70.6% and 69.7%, Spitz's comprehensive gross margins were 72.17%, 69.74% and 58.15%, and Cambrian's comprehensive gross profit margins were 68.19%, 65.38% and 62.39%, respectively.

For Yuncong, if the total revenue continues to decline sharply and the comprehensive gross profit margin further declines, then turning a loss into a profit in 2025 will also become an empty talk.

"Cloud's total revenue and net profit attributable to the parent in the first three quarters are not optimistic, on the one hand, commercialization is not as expected, on the other hand, the outside world's expectations for AI companies are too high, in the future, Yuncong's business performance may still be difficult to boost." Ye Yiming, an investor who has been following the AI field for many years, told Bullet Finance.

2. R&D personnel are reduced

High R&D expenses are an important reason for the cloud's failure to turn a profit from a long time, in fact, not only the cloud, but also the high R&D investment is the "common disease" of AI enterprises, but also the "moat" of AI enterprises.

"Widening the moat as much as possible by increasing R&D investment is still the general consensus of the AI industry, but the huge contrast between high R&D expenditure and the delay in achieving profitability still makes many investors feel like a dragon." Ye Yiming said bluntly.

From 2019 to the first three quarters of 2022, Yuncong's R&D expense ratio was 56.25%, 76.59%, 49.67% and 94.24%, respectively, while SenseTime's R&D expense ratio from 2019 to the first half of 2022 was also as high as 63.30%, 71.21%, 76.89% and 143.82%, respectively.

In the face of high R&D investment, Yuncong still made a compromise.

From 2019 to the first half of 2022, Yuncong's R&D personnel compensation was 224 million yuan, 320 million yuan, 298 million yuan and 113 million yuan, accounting for more than 40% of the R&D expenses in the same period.

In order to throttle, the cloud has been reduced from the number of R&D personnel.

Photogram / Photogram, based on VRF protocol

According to the prospectus and the financial report for the first half of 2022, the number of R&D personnel in Yuncong reached 997 at the end of 2020, which was then reduced to 575 at the end of 2021 and further increased to 609 in the first half of 2022.

In 2021, Yuncong's R&D expenses decreased by 7.6% year-on-year to RMB534 million, and in the first half of 2022, R&D expenses decreased by 3.4% year-on-year to RMB256 million.

However, reducing the number of R&D personnel does not seem to be a safe way to throttle, and while the cloud chooses to reduce R&D personnel, its competitors are actively expanding their R&D teams.

For example, Hikvision had 20,597 R&D personnel at the end of 2020 and expanded to 25,352 by the end of 2021; The number of SenseTime's R&D personnel increased from 2,803 at the end of 2020 to 4,093 in the first half of 2022.

, which is a direct competitor to Cloud, is also expanding its R&D team, from 601 at the end of 2020 to 720 by the end of 2021.

In Ye Yiming's view, AI companies can only return to technology precipitation with high R&D investment, and the loss of R&D personnel from Yuncong may adversely affect its long-term business conditions and market competitiveness.

For example, in terms of the number of invention patents, as of December 31, 2021, the number of invention patents of Hikvision, SenseTime and was 2,243, 2,194 and 210 respectively. In comparison, the number of invention patents in Yuncong is only 133.

R & D personnel are related to the construction of technical barriers of AI enterprises, although reducing the number of R & D personnel reduces part of the cloud from part of the R & D investment, but the overall effect is extremely limited, for AI companies based on high and new technology, by reducing the number of R & D personnel to achieve the purpose of throttling is tantamount to drinking and quenching thirst.

3. Tell the story of the metaverse badly

Open source is obviously more important than throttling.

In addition to the aforementioned benchmarking against OpenAI, Yuncong is trying to catch the "free ride" of the metaverse - laying out on the AIGC (AI generated content) track and creating an AI-driven virtual digital human.

At present, there is a general consensus in the industry that AI technology, as a link between the virtual world and the real world, is one of the key basic technologies for building the metaverse.

In May 2022, at the first Tianfu Metaverse Conference in 2022, Zhou Xi proposed the concept of "AI Genius", he believes that AI Genie is actually an intelligent entity that thinks and acts like a human, helping humans to complete complex tasks and tasks in specific fields or other fields, and has a variety of forms.

In this regard, Yuncong has designed a CWOS human-machine collaborative operating system, which realizes "AI genie thinks and behaves like humans" from three aspects: digitization of the physical world, knowledge recruitment, and accompanying execution, and promotes AI technology to become the basic technology of the metaverse.

However, although the concept of the metaverse has flourished, the cloud has not yet made substantial progress.

On the investor relations platform, Yuncong's secretary said in response to investors' questions that the company will actively pay attention to the progress of the metaverse concept and technology, seek the connection point between AI technology and the metaverse, and explore and expand new areas of AI technology application scenarios in combination with market demand.

Objectively speaking, the metaverse is indeed one of the most suitable scenarios for the landing of AI technology, and is considered to be the most valuable industry in the future.

According to the research report, the size of China's metaverse market in 2022 will be about 42.5 billion yuan, a year-on-year increase of nearly 20%, and it is expected to grow to 126.4 billion yuan by the end of 2027.

At the 6th World Intelligence Conference, Xu Li, CEO of SenseTime, said: "In the next stage, as long as the business model is redefined with the cognition of the true metaverse natives, there is huge room for imagination. ”

Compared with Yuncong, SenseTime has taken the lead in finding a landing scenario that combines the metaverse, such as its cooperation with Centaline Real Estate, the two parties will jointly create a "real estate metaverse", based on SenseTime's SenseMARS Mars mixed reality platform, through AI+MR and other digital technologies, the two parties will jointly develop a series of online "immersive viewing" metaverse applications.

In addition, based on the SenseMARS Mars mixed reality platform, SenseTime has also cooperated with the Palace Press, Xu Beihong Art Museum, and Dunhuang Cultural and Creative to launch digital cultural and creative products that "combine virtual and real", in an attempt to create a metaverse of traditional Chinese culture.

(Photo / "Song Renpi Ma Xu Beihong" (SenseTime Technology Official Weiwei))

However, what needs to be poured cold water is that whether it is Yuncong or SenseTime, its metaverse-related businesses are in the stage of exploration and trial and error, and have not brought any actual revenue.

Ye Yiming analyzed to the "bullet financial view" that due to many restrictions such as technology and landing scenarios, the combination of cloud cong and meta-universe still has major defects, "In the short term, the meta-universe can indeed awaken the vitality of Yuncong and open the upper limit of its commercialization, but in the long run, if the meta-universe product cannot be landed, then the cloud cong can only stay in the stage of hype concept, and it is difficult to tell a new story of the meta-universe." ”

*Note: The names in this article are pseudonyms. The title picture in the article is from: Yuncong Technology official website.

Read on