Recently, DeepSeek, an AI newcomer, has stormed the internet with astonishing "China Speed," rapidly emerging as a phenomenal application. As of February 5, 2025, officials announced that its daily active users have surpassed 40 million, with monthly active users across all platforms reaching 170 million, setting a new record for this milestone. DeepSeek's popularity not only shines brightly but also invigorates the entire AI industry chain.

Currently, China's three major telecommunications operators—China Mobile, China Unicom, and China Telecom—have fully integrated the DeepSeek large model, providing exclusive computing power solutions and supporting environments to unleash the full potential of domestic large models. Meanwhile, mainstream cloud service providers like Huawei Cloud, Tencent Cloud, Alibaba Cloud, and Baidu Cloud have successively launched DeepSeek's large model services on their platforms, further expanding its application scenarios.

Facing this opportunity, leading domestic cloud computing companies are rushing to integrate DeepSeek, potentially driving rapid growth in cloud service providers' revenue from computing power leasing and AI services. Furthermore, DeepSeek's innovative low-cost technology for open-source models will reduce barriers to application development, accelerate the prosperity of application-side developments, and thus drive a sustained increase in cloud computing power demand.

Given DeepSeek's significant benefits and crucial position in the cloud computing field, it is essential to conduct in-depth research and understanding of this emerging technology.

01. What Changes is DeepSeek Bringing to the Cloud Computing Industry?

With its remarkable advantages, the DeepSeek model is profoundly impacting the cloud computing industry. Its performance rivals the world's top models, featuring humanization, low cost, and open-source characteristics, collectively shaping a new landscape in cloud computing and large models.

In terms of cost, DeepSeek significantly reduces training and inference costs, leading to a fundamental shift in the cloud computing business model. Previously, closed-source high-profit models were the norm, but DeepSeek's open-source low-profit model has brought new vitality to the market. This change not only lowers barriers for customers but also prompts cloud vendors to re-examine their business models and seek survival and thriving strategies in the new market environment.

Deployment flexibility is another significant change brought by DeepSeek. Since it can be deployed on the edge side, the inference services provided by cloud vendors allow customers to easily switch between different vendors. This low-cost switching mechanism, such as Tencent's limited-time free strategy to attract customers, further intensifies market competition and urges cloud vendors to continuously improve their service quality and technological level.

In terms of computing power demand, DeepSeek's emergence allows domestic vendors to significantly narrow the gap with overseas counterparts in a short period. While training may be affected to some extent, the demand for inference has shown an explosive growth trend. It is estimated that by the end of 2026, the demand for inference may increase a hundredfold compared to the beginning of 2025, with domestic production also experiencing significant growth, providing new development opportunities for cloud vendors.

DeepSeek's thorough open-source nature has profoundly impacted cloud vendors' business models. Customers do not need to pay any fees for self-deployment using DeepSeek models, prompting cloud vendors to adjust their charging logic. They now generate revenue through API services, leasing computing power, hybrid clouds, or private deployments. Under this new business model, cloud vendors need to deploy quickly, support higher concurrency, and lower latency to meet customer needs.

For customer self-training models, cloud vendors also need to provide more support, including data and storage requirements after algorithm optimization. Whoever excels in this area will gain an advantage in the hybrid cloud and government cloud sectors. Therefore, cloud vendors are continuously increasing investment in technology research and development and customer service to enhance their competitiveness.

02. Analysis of the Cloud Computing Industry Chain

Currently, the cloud computing industry chain is receiving renewed attention from relevant institutions. The author has compiled a list of relevant companies for investors' reference:

Three Major Operator Clouds:

As the leading force of the national team, China Telecom's Tianyi Cloud, China Mobile's Mobile Cloud, and China Unicom's Unicom Cloud leverage their network resources and government and enterprise customer advantages to dominate the government cloud market. They focus on data center construction, industry solutions, and hybrid cloud services through ecological cooperation. Among them, Tianyi Cloud collaborates with Sundata A (cloud-data integrated machine), Chena Anke (smart city cloud platform), etc.; Mobile Cloud partners with UCloud (hybrid cloud technology), Capital Online (edge computing nodes), etc.; and Unicom Cloud collaborates with Wangsu Technology (content distribution network CDN acceleration), Donghua Software (government cloud integration), etc.

Internet Giant Clouds:

Huawei Cloud: With the 'Kunpeng + Ascend' ecosystem as its core, it focuses on government and enterprise, finance, and manufacturing sectors. Its partner enterprises cover the entire chain from hardware to applications, including Tuowei Information (hardware and software integrated partner), Changshan Beiming, Zhongyi Technology (database migration), etc.

Alibaba Cloud: Relying on the Flying Apsaras operating system and DAMO Academy technology, it focuses on e-commerce, finance, and industrial internet. Its core partners include Inspur Information (high-performance computing devices), Hangzhou Iron and Steel (low-carbon data centers), etc., while also collaborating with Yonyou Network (ERP and CRM cloud migration) at the SaaS level.

Tencent Cloud: Leveraging its differentiated advantages in the WeChat ecosystem, audio-video technology, and game cloud, it cooperates with Baoxin Software (data center services), Donghua Software (Tencent large model partner), and other enterprises, focusing on scenario implementation.

Baidu Cloud: With AI large models and autonomous driving clouds as its core, it focuses on technology integration with enterprises such as ThoughtWorks (industry large models) and Jiafei Xunhong (industrial AI quality inspection).

JD Cloud, ByteDance Cloud, Kingsoft Cloud: Breaking through in vertical tracks, they cooperate with Wajinke (retail supply chain cloud), Runze Technology (ultra-large-scale IDC), Kingsoft Office (collaborative office cloud), etc., respectively.

Cloud Computing Service Providers:

Independent cloud service providers form differentiated competitiveness through niche technologies, regional markets, or industry know-how. For example, Qi'anxin (AI-driven cloud-native security protection needs), Qingcloud Technology (hybrid cloud management platform), Capital Online (global node deployment), etc., exhibit strong professional capabilities in fields such as network security clouds, edge computing, and cross-border cloud services.

- End -