In 2023, Cloudwalk Technology will still lose money.
Its performance shows that Yuncong Technology will lose 644 million yuan in 2023, and the loss in the first quarter of 2024 will further expand by 11.62% year-on-year, with a loss of 160 million yuan.
Not only has it not achieved profitability, but it is also constantly facing the challenge of competition from artificial intelligence companies and the promotion of artificial intelligence empowerment transformation of products by large equipment manufacturers and Internet companies.
At the same time, other AI unicorns are also facing the same situation: SenseTime will lose 6.495 billion yuan in 2023, an increase of 6.6% over the same period in 2022. Yuntian Lifei will lose 381 million yuan in 2023 and 138 million yuan in the first quarter of 2024, with the loss widening by 52.67%.
It can be seen that AI models are reshaping all walks of life, and the first batch to be rectified are its developers
01. Industry light, or industry light?
After Yuncong disclosed its 2023 annual report, although it lost money, its operating cash flow situation improved.
The core human-machine collaborative operating system and pan-AI business scale growth has accelerated, the hematopoietic capacity has been significantly enhanced, and the fundamentals of operation have also continued to improve, laying a solid foundation for the company's new round of development.
By 2024, Yuncong's business chassis will begin to be more stable, and with the release of favorable policies such as new quality productivity, Yuncong Technology is predicted by the industry to accelerate from the recovery period to the comprehensive development period, and the market share and performance are expected to accelerate the return.
It was also from this year that Cloudwalk Technology joined the "100 Model War".
Not long ago, Yuncong Technology's application for a private placement plan was accepted by the Science and Technology Innovation Board, which means that it will continue to raise funds, planning to raise more than 3.6 billion yuan, before which its IPO raised a total of 1.728 billion yuan, which is just less than a year before its IPO.
Throwing out a large fundraising plan again, it is enough to see that Yuncong's R&D investment has made its funding gap even larger, but what is even more incredible is that it is clear that Yuncong's technology stock price has risen by more than 130% since 2023, which can be regarded as eating the dividends of this AI wave, why is it still "short of money"?
Although Yuncong is one of the few AI companies that have successfully gone public, its high-flying strategy is repeatedly plunging it into anxiety.
It believes that the research and development of large models has many necessities, including providing efficient human-machine collaborative operating systems and industry solutions, boosting the process of artificial intelligence industrialization and intelligent transformation and upgrading of various industries, among which the construction of large model research and development projects will greatly enhance its ability to recognize the human-machine collaborative operating system.
In the face of investors' questions about when to make a profit, Cloudwalk said that profitability is not a strategic focus this year, but it will reduce losses by optimizing the cost structure, achieve breakeven as soon as possible, and lay the foundation for stable profitability in the future.
From the perspective of specific products, Yuncong mainly relies on providing human-machine collaborative operating systems and application software and artificial intelligence solutions for government and enterprise customers such as governments, public security, banks, airports and other large enterprises, as well as manufacturers or integrators directly facing medium and large end customers, so as to obtain sales revenue.
The variables will appear in 2022, and the ChatGPT released by OpenAI effectively solves the previous problem of low marginal utility of artificial intelligence, so in the face of the new technical route, Yuncong officially launched its self-developed "calm large model", focusing more on the application of the B-end industry vertical market.
It is worth noting that in the past, Cloudwalk was mainly a business on the To G side, and was forced to reduce the proportion of the To G side due to environmental impact.
Because of the expansion to the broader B-end industry, Cloudwalk Technology's revenue growth in the first quarter is quite contrasting, whether it is the development of the Calm Large Model itself or the joint launch of the Calm Large Model Training and Push All-in-One product with Huawei Ascend, all reflect two facts:
First, the large model has the infrastructure value that should not be underestimated, and secondly, Yuncong has been deeply involved in the AI industry for many years, and has long mastered the key to transforming technology into productivity tools.
Capitalists only care about when they can make a profit, but the existence of Yuncong is like a beacon that illuminates the future corner of the large model dispute.
02. Small models are difficult to make money, and large models are too expensive
The future of Cloudwalk Technology is not a smooth one.
Although its business prospects are promising, but also accompanied by a large number of strong competitors, in just half a year, the domestic model accelerated the development of iFLYTEK's "Spark", Tencent's "Mixed Yuan", Baidu's "Wenxin Yiyan" have come out of the laboratory, facing the public to be tested.
Bill Gates once said: AI has the ability to change the world, and the most important thing for AI in this decade, ChatGPT confirms Bill Gates's view, and artificial intelligence applications are opening up a new situation.
iFLYTEK, which has played the slogan of "1024 benchmarking ChatGPT", and SenseTime, which compares its new model InternLM-123B with Meta llama2, have all focused on the large model of 100 billion parameters.
It's just the landing of multiple scenarios, how to better contribute to economic development and enterprise efficiency, which is a difficult problem that "cloud servants" must not face.
On the road of AI large models, only continuous research and development is king.
Major players have long understood the rules of the industry: only by making AI truly a productivity tool can enterprises live better and enter a virtuous cycle of technology and business.
In the process of this transformation, only by accumulating enough basic knowhow skills for all walks of life, continuously expanding the ecosystem, and exploring solutions to productivity problems, can the market accelerate the value of AI and large models.
This is the rigid rule of To B, and To C is also the choice of some enterprises, such as Kimi on the dark side of the moon, which actually chooses to face C-end users.
The founder of Tezan, a technology unicorn company focusing on the field of AIGC, once said: Now large-scale model entrepreneurship is very specific, either to help enterprises make money within a year, or to help customers increase efficiency within 3 months, in essence, technology must serve these two purposes.
Tencent Technology has previously reported that China's technology industry's attitude towards large models has split into two camps: one group is more confident in technology, and they believe that if they do not pursue "bigger and stronger AI capabilities", they will soon reduce the dimension and crush the existing base and moat;
The other group believes that as long as "sufficient AI capabilities" are put into business scenarios that can be quickly monetized, they can use the huge and unique data of the Chinese market to build barriers.
Of course, the former must be the lifelong pursuit of all large-scale model companies, but the latter is the real status quo of more Chinese large-scale model entrepreneurs.
It is difficult for small models to make money, and large models are too expensive, which is the result of many factors such as "expensive" large models and more cautious capital venture capital.
According to the "2023 Artificial Intelligence (AI) Industry Status Report", in 2023, the number of investment and financing in China's AI field will be about 232, a year-on-year decrease of 38%; The total amount raised was about $2 billion, down 70% year-on-year. Especially in the first quarter of the past year, both the amount of financing and the amount of financing have hit the lowest level in five years, and the "cooling" is particularly obvious.
At this stage, the domestic model has been in a world-leading position at the regulatory level, and it is more necessary to encourage development, so that development and regulatory security can coexist, and more resources should be given to young scholars and entrepreneurs to give them room for trial and error.
Yuncong has been in the industry for many years, and the next start-up companies should not consider the distant dream of AGI, but focus on the input-output ratio to lay a solid foundation for the industry.
Reference:
The difficulty of large-scale model entrepreneurship: it is difficult for small models to make money, and large models are too expensive, and the amount of investment and financing shrank by seventy percent last year-Hexun.com
The current situation of investment and financing of China's AI large model: the secondary market is stupid, and the primary market is deserted - the new intellectual world of AI
With a revenue of 54 million in the first quarter and a loss of 160 million, when will Cloudwalk Technology turn losses into profits? - Yang Jianyong
2023 State of the Artificial Intelligence (AI) Industry Report