Article | AI Relativity Theory

As the former "first mixed cloud stock", QingCloud Technology also had its glory days. In 2021, it was listed on the STAR Market of the Shanghai Stock Exchange, with a market value once reaching RMB 4.034 billion. However, it has now halved to RMB 1.733 billion.

What has happened to QingCloud Technology over the years? Despite being in one of the hottest sectors of the era, it has struggled to maintain its former glory. Is it the bursting of an era's bubble, or is it a strategic misstep by QingCloud Technology?

Not long ago, QingCloud Technology announced the launch of upgraded products and services, industry and scenario solutions, as well as an ecological strategy. Centered around AI intelligent computing platforms, AI computing clouds, AI intelligent computing all-in-ones, and other products and services, the company aims to enhance overall computing efficiency and optimize the allocation and utilization of computing resources. QingCloud Technology hopes to provide flexible and diverse heterogeneous computing solutions for the entire industry and various scenarios.

From mixed cloud services to a focus on intelligent computing power, QingCloud Technology has been adjusting its business focus in response to changing market demands. However, facing the increasingly competitive AI industry, QingCloud Technology still finds it difficult to break through in a market dominated by giants, finding itself stuck in a dilemma of not being able to move up or down.

QingCloud's Dilemma: Consecutive Losses, Unable to Break the Stalemate

According to the latest 2024 half-year report, QingCloud Technology is still in a state of loss, with a net loss of RMB 38 million, a narrowing compared to the net loss of RMB 82 million in the same period last year. However, what's worse is that revenue also declined in the first half, with revenue of RMB 144 million, a year-on-year decrease of 17.51%.

This is not a short-term dilemma for QingCloud Technology. From 2019 to 2023, the company has been incurring losses, with net losses of RMB 190 million, RMB 163 million, RMB 283 million, RMB 244 million, and RMB 170 million, respectively. The company has continuously "burned money" each year, and its listing has not reversed the trend of losses.

Of course, losses are not the crux of QingCloud Technology's dilemma. After all, being in an emerging technology industry, many companies face the situation of "burning money" to gain market share. However, the problem lies in the fact that QingCloud Technology has not gained corresponding market returns, making it difficult to see its ability to break the stalemate in the future.

From 2019 to 2023, QingCloud Technology's revenue was RMB 377 million, RMB 429 million, RMB 424 million, RMB 305 million, and RMB 336 million, respectively. Not only did revenue decline significantly in two of these years, but considering the first half of this year, it seems somewhat embarrassing for QingCloud Technology to return to its past peak.

At this critical stage, QingCloud Technology seems to hope to stabilize the company's loss state by reducing R&D investment. In the first half of this year, QingCloud Technology's R&D investment was RMB 35 million, a decrease of 24.2% compared to the same period in 2023. Looking further back, the R&D investment in the same period in 2023 was RMB 46 million, a decrease of 37.47% compared to RMB 74 million in the same period in 2022.

With such a reduction in R&D investment, corresponding internal layoffs are inevitable. In fact, QingCloud Technology is indeed optimizing its R&D team, with the number of R&D personnel decreasing from 346 as of June 30, 2022, to 263 as of June 30, 2023, a reduction of approximately 23.99%.

While the reduction in R&D investment in the short term helps improve financial statements, it may negatively impact the company's technological innovation capabilities and market competitiveness in the long run. Moreover, QingCloud Technology is in a fiercely competitive cutting-edge technology sector where technological innovation and R&D are of utmost importance. QingCloud Technology's "cutting" of R&D investment and personnel seems to have reached a stage of desperation.

But is this approach really about "survival" or "death"? It's hard to say. Based on the situation in the first half of this year, it seems that QingCloud Technology is still unable to break the stalemate.

Intelligent Computing Power, Difficult to Solve QingCloud's Profitability Dilemma

In recent years, QingCloud Technology has been hoping to focus more on business areas that can bring higher returns with limited resources. For QingCloud Technology, this is an ideal state, where by reducing investment in non-core projects, more funds and human resources can be invested in its most competitive products and services, with the aim of increasing market share, enhancing customer loyalty, and ultimately improving financial performance.

At the same time, QingCloud Technology has emphasized AI computing power services as an important new growth point, believing that AI has opened up a new incremental market for cloud computing and sees it as an important opportunity for the next decade. But can intelligent computing power really solve QingCloud Technology's dilemma?

QingCloud Technology's revenue from AI computing power services was RMB 26 million, a year-on-year decrease of 10.98%, while revenue from AI intelligent computing platforms and strategic software was RMB 11 million, a year-on-year decrease of 20.29%. It is evident that the new growth point has not become an explosion point for QingCloud Technology, and the performance of intelligent computing power-related businesses remains weak.

Why is this the case? Intelligent computing power is not as easy as QingCloud Technology imagined.

Although both hybrid clouds and intelligent computing power belong to the field of cloud computing, they are entirely different concepts with different focuses, serving different application scenarios and technical requirements.

Specifically, intelligent computing power focuses on computing power itself, especially the ability to efficiently handle compute-intensive tasks. In contrast, hybrid clouds focus on how to allocate and manage computing resources across different types of cloud environments to meet specific enterprise needs. Both can coexist in an organization's IT strategy. Intelligent computing power can be implemented in any part of a hybrid cloud architecture, whether it's a private cloud or a public cloud, and systems with high intelligent computing power can be deployed to support various advanced computing tasks.

This means that to excel in intelligent computing power, QingCloud Technology must increase investment, whether at the R&D level or the infrastructure level, requiring continued "money-burning".

On the one hand, intelligent computing power involves a large amount of computing resources, especially in the current process of large model training and inference, where the demand for computing power is extremely high. QingCloud Technology needs to provide high-performance computing devices such as GPUs and other specialized accelerators. On the other hand, it is also necessary to build corresponding infrastructure to support the efficient operation of these devices. This includes not only hardware facilities but also a powerful software scheduling platform to manage these computing resources, enabling them to be provided to customers as a service.

This is a rather complex systematic solution. For QingCloud Technology, which is currently incurring losses, it seems difficult to have sufficient resources and capital to invest in-depth in intelligent computing power.

In particular, the customer demands currently facing intelligent computing power are also relatively complex, often requiring specialized and customized solutions that are difficult to meet with a standardized solution. For example, enterprise customers face multiple challenges when applying AI technology, including how to integrate AI technology into their existing business processes and how to manage and optimize the storage and processing of large amounts of data. This requires QingCloud Technology to have a deep understanding of the business needs of different industries and provide customized solutions to meet these needs.

Objectively speaking, in the field of intelligent computing power, QingCloud Technology may not be able to recreate the glory of being the "first mixed cloud stock" and can only serve as an introduction to the next business story.

Conclusion

QingCloud Technology's dilemma stems from both internal and external factors.

1. In China's cloud computing market, QingCloud Technology not only has to compete with domestic large cloud service providers such as Alibaba Cloud, Tencent Cloud, and Huawei Cloud but also faces competition from state-owned clouds such as China Telecom's Tianyi Cloud, China Unicom Cloud, and China Mobile Cloud. These competitors have abundant resources, high market share, strong brand influence, and a vast commercial ecosystem for expansion, making it difficult for QingCloud Technology to compete head-on.

2. QingCloud Technology has deepened its focus on the hybrid cloud field, mainly targeting traditional enterprise users. These customers usually have high requirements for customized services and strict requirements for security and stability. To meet these demands, QingCloud Technology must invest more resources in customized development and service support, leading to high costs that are difficult to sustain.

3. Whether in the hybrid cloud sector or the field of intelligent computing power, to maintain a leading position in technology and solutions, QingCloud Technology needs to continuously increase investment in R&D. However, the high R&D costs are difficult to translate into revenue growth of the same scale in the short term.

In summary, QingCloud Technology's path is not easy. Simply turning losses into profits is already a challenging task for the former "first mixed cloud stock," let alone other matters that have to be put on hold.

*All images in this article are sourced from the internet