AI companies are listed on the market. The fourth paradigm has re -battle Hong Kong stocks. The losses in the first half of the year have narrowed half of the year, saving money for the epitome of the winter industry

China Times (chinatimes.net.cn) reporter Yan Xiaohan Lu Xiao Beijing Beijing report

Although since last year, Shangtang Technology, Yun, and many AI companies such as technology, innovation and wisdom have been successfully listed after twists and turns, but most of the AI ​​companies have gone up.

On September 5th, the fourth paradigm of artificial intelligence companies submitted a prospectus to Hong Kong stocks for the third time, and updated its financial data for the whole year of 2021 and the first half of 2022. In the first half of this year, the fourth paradigm revenue was 1.058 billion yuan, an increase of 34%year -on -year; the net loss of the loss was 200 million yuan, which was about 20%year -on -year; half.

Compared with the hustle and bustle of the AI ​​track a few years ago, the AI ​​industry has entered the cold winter in the past two years. When the key indicators of the outside world pay attention to, almost no players can give the optimal solution. But one of the changes to be seen is that in the first half of this year, cautious investment and reduction of losses have become a common choice for most companies.

The proportion of R & D investment declines

Founded in 2014, the fourth paradigm listing process did not go smoothly. In August last year, the fourth paradigm submitted the prospectus to the Hong Kong Stock Exchange for the first time, and it failed in February this year. Subsequently, it submitted the prospectus again within 10 days, and the status of its IPO application state after half a year later became invalid.

For most AI companies that are not good for performance, the success of the success of the capital market means that it is necessary to accept more review and challenges.

From the perspective of the revenue performance of the fourth paradigm in 2021, its revenue still maintained a high increase of 114%to 2 billion yuan. Adjusting net losses also expanded by 43%to 560 million yuan.

The main reason for losses is the old problem of most AI companies: long -term continuous research and development investment, and the amount is too large. In the whole year of 2021, the fourth paradigm invested 1.25 billion yuan in R & D, which was more than 260 million yuan more than the sum of R & D investment in the two years of 2019 and 2020.

Although compared with the growth rate of more than double the past three years, the fourth paradigm revenue in the first half of this year has slowed significantly, its losses have also decreased significantly. The main reason is that the pay -based salary expenditure is reduced.

One change that needs to be seen is that although AI companies generally need to keep "burning money" to maintain their technical advantages to gain a foothold. But in the first half of this year, the proportion of R & D investment in the fourth paradigm fell from 73.4%in the first half of last year to 52.7%. In addition, the proportion of sales and marketing spending in the first half of the year also declined.

The fourth paradigm stated in the prospectus that it is especially important to improve operating efficiency by controlling operating expenses. As the business scale increases, it is expected to produce a significant operating scale effect and achieve structural cost savings for effective competition. "Looking forward to the future, we will continue to be cautious about the research and development resources of the platform and infrastructure."

It is not the fourth paradigm that reduces various investment such as R & D. The proportion of R & D investment in the first half of this year in the first half of this year. In the first half of this year, its R & D investment accounted for 48.73%; the proportion of the same period last year was 76.51%.

Zhang Xiaorong, the dean of the In -depth Science and Technology Research Institute, analyzed the reporter of the Huaxia Times that R & D investment has greatly affected the competitiveness of enterprises. A large part of R & D investment is personnel expenses. The entire R & D team and R & D personnel are relatively high at the industry. The cost is very expensive, and the number of personnel is huge, which leads to the high cost of research and development. The second is to purchase equipment, purchase software and hardware, and purchase patents. If the enterprise needs to reduce costs, the costs of these two parts will be compressed and the cost of compression personnel has the greatest impact on the enterprise.

"But this is not possible for the current AI companies. The entire AI industry has entered the trough period or the cold winter stage. Even if the money -making companies have felt the cold, the colds that these non -profit companies feel more obvious and more urgent. Therefore, they will take the initiative to take measures to open source and reduce costs. First, go through winter and then seek rapid development. "Zhang Xiaorong said.

On September 8th, the reporter of the Huaxia Times contacted the fourth paradigm on the decline in the proportion of R & D investment. It stated that the company was in a silent period and was inconvenient to reply.

A number of AI companies have narrowed their losses in the first half of the year

Artificial intelligence is still in the initial stage, and its losses are the norm. The reason for the difficulty of profit is that the technology is not mature enough, and the operating costs are high. But in the first half of this year, most of the listed AI companies have dawned the narrowing of losses.

In the first half of this year, the clouds from the science and technology board of the Science and Technology Board have reduced sales, management and research and development costs in the first half of the year. Net losses have been narrowed by 11%to 325 million yuan. The loss was narrowed by 77.9%year -on -year, mainly due to increased income and decrease in share payment; innovative wisdom was narrowed by 3%in the first half of the year.

Zhang Xiaorong's analysis of the "Huaxia Times" believes that the narrowing loss of many AI companies in the first half of the year was mainly affected by the external environment. At present, the market consumption is sluggish. Customers in the AI ​​industry are more TOB and TOG, and the company has more government orders. In the past two years, the epidemic has been repeated, and the budget of the government spent on AI has also decreased year by year. Therefore, companies should cut costs and reduce their investment budget or investment scale. "The project of spending money is less. Although it is not very much earned, the loss is narrowed a bit." It should be mentioned that different AI companies have different layout tracks. For example, the "AI Four Little Dragons" founded around 2010, and the Galing deep pupils established in 2013, started in the field of computer vision. They were favored by capital because AlphaGo defeated the human world Go champion to the spotlight in 2016.

The fourth paradigm layout is decision -making artificial intelligence. For example, by summing up the laws of massive data, it helps retailers to predict sales to formulate precision marketing strategies. On the platform -centric decision -making artificial intelligence track, the fourth paradigm also shall also have wrists with Internet factories such as Baidu, Tencent, Huawei, Ali.

However, Zhang Xiaorong believes that the market size of AI companies is currently limited, and the industry is still in the initial stage. Everyone belongs to free competition. The degree of competition looks fierce as a whole, but it is not desperate to be desperate. Everyone does not have a direct competition in the direct sense, but more reflected in product ideas and solutions, such as whether product technology is stronger and whether the solution is more cost -effective.

He said that the obstacles to landing also limited the growth rate of artificial intelligence enterprises. "The reason why AI's development progress is very slow is that to adjust a parameter, enterprises need a large amount of data deduction and training, so that the model will move towards and mature. It can be used for limited scenes, and there are often some unsatisfactory places. This is also the dilemma facing most AI companies. "

Editor -in -chief: Editor -in -chief Huang Xingli: Han Feng

- END -