Recently, Kingsoft Cloud Holding Limited (hereinafter referred to as "Kingsoft Cloud"), which was highly anticipated by Lei Jun, announced its performance for 2023. Following a decline in revenue in 2022, Kingsoft Cloud's revenue once again fell in 2023, with a loss exceeding 2 billion yuan.
In the past year, facing the siege from cloud computing giants such as Alibaba Cloud and Huawei Cloud, Kingsoft Cloud had to initiate a price war, but the actual effect was not ideal. In the capital market, Kingsoft Cloud's stock price plummeted by over 95% from its peak, leaving investors empty-handed.
Under the siege of giants, Kingsoft Cloud is pinning its hopes on AI. But can Kingsoft Cloud successfully break through with the support of AI?
In 2012, Kingsoft Software divested its KuaiPan business to its newly established subsidiary, Kingsoft Cloud, and officially entered the cloud computing field. Backed by Lei Jun's aura and the booming cloud computing industry at the time, Kingsoft Cloud secured over $100 million in financing from its parent company Kingsoft Software, Xiaomi Group, and IDG Capital within a few years of its establishment.
With the support of capital and the booming cloud computing industry, Kingsoft Cloud's revenue grew rapidly. In 2018, Kingsoft Cloud's revenue was 2.218 billion yuan, which increased to 9.061 billion yuan in 2021, with a compound annual growth rate of nearly 60%.
In 2020, Kingsoft Cloud had its shining moment when the company was listed on the US stock market at a price of $17 per share. On the first day of listing, Kingsoft Cloud's stock price surged by over 20%. In 2021, Kingsoft Cloud's stock price once reached over $74 per share, and its market value exceeded $17.6 billion at one point.
However, the good times didn't last. As the enthusiasm for cloud computing in the capital market waned and Kingsoft Cloud's revenue began to decline, its stock price began to plummet.
As of April 12, 2024, Kingsoft Cloud's stock price closed at $2.70 per share, down over 95% from its peak, and down over 84% from its IPO price. This also means heavy losses for investors like Lei Jun.
The external cause of Kingsoft Cloud's stock price collapse is the abandonment by capital, while the internal cause is the decline in the company's revenue. In 2021, Kingsoft Cloud's revenue reached a record high of 9.061 billion yuan, but it has not exceeded 9 billion yuan since then.
In 2022, Kingsoft Cloud's revenue was 8.108 billion yuan, a year-on-year decrease of 9.72%, and the company's loss expanded from 1.589 billion yuan in 2021 to 2.658 billion yuan. In 2023, Kingsoft Cloud achieved revenue of 7.047 billion yuan, a year-on-year decrease of 13.85%, with a net loss of 2.176 billion yuan.
In terms of revenue, Kingsoft Cloud's revenue is mainly divided into public cloud services and industry cloud services. In 2023, Kingsoft Cloud's revenue from public cloud services was 4.382 billion yuan, accounting for 62.18% of its total revenue, making it the largest business segment for Kingsoft Cloud.
The decline in revenue from public cloud services is also an important reason for the overall revenue decline of Kingsoft Cloud. In 2022, the company's revenue from public cloud services was 5.36 billion yuan, which decreased to 4.382 billion yuan in 2023.
The company's public cloud services mainly consist of cloud computing, cloud storage, and cloud distribution, with cloud distribution accounting for the highest proportion of revenue in the public cloud. Cloud distribution relies on Content Delivery Networks (CDNs) to ensure quick loading of images and videos.
Essentially, CDNs face almost no technical barriers, and having more CDN nodes to ensure user browsing speed is the only competitive advantage for relevant enterprises.
However, deploying CDN nodes requires significant financial support. Data shows that in 2023, Alibaba Cloud had 3,200 global nodes, Tencent Cloud had over 2,800, and Huawei Cloud had over 2,800 as well. In contrast, Golden Mountain Cloud has significantly fewer nodes, with only 1,700 globally.
Compared to large players like Alibaba Cloud and Tencent Cloud, Golden Mountain Cloud is clearly at a significant disadvantage in terms of funding and operational scale. It's not surprising that Golden Mountain Cloud's revenue is declining under pressure from numerous large cloud service providers.
The cloud computing industry is witnessing intensified competition.
According to data from IDC, in the first half of 2023, the IaaS+PaaS market size of China's public cloud services was $14.58 billion, with a year-on-year growth of 15.9%, marking the lowest growth rate in nearly three years.
To attract more customers amid slowing industry growth, major cloud service providers have begun intensifying competition. In April 2023, Alibaba Cloud initiated the largest-scale market price reduction, slashing core product prices by 15% to 50%, with storage product prices dropping by up to 50%.
Subsequently, Tencent Cloud, Mobile Cloud, and Tianyi Cloud, among other major cloud service providers, also followed suit with price cuts. Tianyi Cloud even advertised products at discounts of up to 90%.
In response to the intensified competition, Golden Mountain Cloud also reduced prices, with its CDN and cloud database product Reds experiencing price cuts of up to 50%.
Interestingly, despite the price war, private enterprises like Alibaba Cloud and Golden Mountain Cloud did not benefit; instead, the three major telecom operators swiftly expanded their cloud services within the industry.
In 2023, the performance of cloud computing businesses of the three major telecom operators was impressive, with Tianyi Cloud achieving a revenue of 97.2 billion yuan, a year-on-year growth of 67.9%; Mobile Cloud achieving a revenue of 83.3 billion yuan, a year-on-year growth of 65.6%; and Unicom Cloud achieving a revenue of 51 billion yuan, a year-on-year growth of 41.6%.
In contrast, Alibaba Cloud's revenue in 2023 was 105.396 billion yuan, a slight year-on-year increase of 3%. Golden Mountain Cloud fared even worse, with its revenue declining by 13.85% year-on-year in 2023. This indicates that Golden Mountain Cloud's strategy of reducing prices to attract customers did not yield the desired results, raising concerns among investors about its competitiveness.
Unlike the high-speed growth of the past decade, the cloud computing industry is gradually entering a period of stable growth. According to the "White Paper on the Development of China's Cloud Computing Industry," it is predicted that by 2023, the cloud adoption rate of the Chinese government and enterprises will reach 60%, further slowing industry growth.
At the government and state-owned enterprise levels, the three major telecom operators have a natural advantage, with more data centers and higher stability. Private enterprises like Golden Mountain Cloud naturally face disadvantages at the government and state-owned enterprise levels, making it difficult for them to make breakthroughs.
In the enterprise sector, due to the homogeneity of services provided by major cloud service providers, customers are unlikely to switch providers unless there is a major outage. Considering the potential security risks involved in migrating data between different cloud servers, customers are generally reluctant to switch service providers, leading to higher customer retention rates for cloud service providers. This means that even with a price war, Golden Mountain Cloud is unlikely to attract many new customers.
Furthermore, among enterprises that have not yet adopted cloud computing, many are not very interested in it, especially small and medium-sized enterprises in the manufacturing industry, which may not purchase cloud computing services due to cost concerns.
For Golden Mountain Cloud, the current situation is undoubtedly challenging. In a market with limited growth potential, and with pressure from major players like the three major telecom operators and Alibaba Cloud, expanding outward is proving to be extremely difficult for Golden Mountain Cloud.
Moreover, under the onslaught of price wars initiated by major cloud service providers, Golden Mountain Cloud has lost ground continuously, as evidenced by its declining revenue. Facing this predicament, Golden Mountain Cloud has stated that it will strengthen cooperation with Xiaomi Group and Kingsoft Corporation's affiliated customers. With the support of affiliated enterprises, Golden Mountain Cloud may still have a chance to turn things around.
Due to its poor operational performance, Golden Mountain Cloud is actively seeking transformation.
In 2023, with the popularity of ChatGPT, major cloud service providers accelerated their AI layout. For example, under Baidu Cloud's "Cloud and Intelligence Integration" strategy, its product lineup has seen significant improvements, and Alibaba Cloud has begun integrating the Tongyicunwen AI mega model.
Currently, major cloud service providers are intensifying their efforts in AI mega model research and development, marking the beginning of a grand battle of mega models. Public data shows that there are over 200 AI mega models in China so far.
Golden Mountain Cloud is no exception. As early as 2023, it released three generative artificial intelligence mega model images. In March 2024, global growth consulting firm Frost & Sullivan published the "2024 China Mega Model Capability Evaluation," which conducted authoritative evaluations of 15 mainstream mega models in China.
Baidu's Wenyanyixin and Tencent's Huanyuan models belong to the first tier, while Golden Mountain Cloud's related mega model products did not even enter the evaluation scope.
Whether in cloud computing or mega model development, a large amount of funding is needed as the foundation for research and development. Previously, it was reported by The Paper that training a new mega model initially costs 200 million to 300 million yuan. Due to the rise in computing power prices, it may now cost up to 500 million yuan to train a new mega model. According to Guosheng Securities' previous forecast, the cost of training GPT-3 in China is approximately $1.4 million per run, while the training cost for large language models is even higher, ranging from $2 million to $12 million.
As a medium-sized cloud service provider, Kingsoft Cloud is not financially robust. As of the end of 2023, Kingsoft Cloud's cash and cash equivalents totaled ¥2.255 billion.
It's worth noting that from 2022 to 2023, Kingsoft Cloud incurred operating losses of ¥2.658 billion and ¥2.176 billion respectively, indicating that the company's existing funds are only sufficient to cover normal losses for one year.
If Kingsoft Cloud increases its investment in AI large model research and development, its losses will inevitably expand further. In fact, due to declining revenue in the past two years, Kingsoft Cloud has not only failed to increase its research and development efforts, but has also reduced R&D expenditures. From 2021 to 2023, Kingsoft Cloud's R&D expenditures decreased year by year, from ¥1.044 billion to ¥0.971 billion to ¥0.785 billion.
Continuously investing in research and development is the foundation for the continuous iteration and optimization of large models. However, based on Kingsoft Cloud's monetary funds and R&D investment, the company does not have the ability to sustain continuous iteration.
For Kingsoft Cloud, its traditional cloud computing business is struggling under the siege of large cloud service providers. Wanting to leverage AI for advancement but constrained by insufficient funds, perhaps as Kingsoft Cloud stated, only by continuously embracing Kingsoft Software and Xiaomi Group can Kingsoft Cloud gain a lifeline in the ruthless industry competition.
Editor: Li Yufei
Reviewer: Songwen
