Amid the wave of digital transformation, artificial intelligence (AI) is emerging as a core driver of social progress. Intelligent Computing Center As a critical infrastructure for AI technology applications, it shoulders the heavy responsibility of data storage, processing, and analysis, serving as the cornerstone of AI innovation and deployment.

Therefore, The construction of intelligent computing centers has become a critical component of China’s digital transformation. In February of this year, the State-owned Assets Supervision and Administration Commission of the State Council convened a special symposium on artificial intelligence for central enterprises titled “AI Empowers Industrial Renewal,” at which it called for strengthening the foundational infrastructure for AI development, concentrating key resources on the areas with the greatest need and the strongest competitive advantages, and accelerating the construction of a number of intelligent computing centers.

Under the concrete actions and call-to-action of the national government, substantial private capital has also flocked to the construction of intelligent computing infrastructure. This includes both companies in the upstream and downstream segments of the original computing power industry, as well as numerous cross-industry firms that are optimistic about the development of the intelligent computing industry and have made substantial investments.

Notably, A-share listed companies have been the most prominent in this regard. On the one hand, these firms boast substantial financial resources and frequently commit investment tranches of RMB 1 billion or even several billion to develop related sectors. On the other hand, their business activities must balance market-capitalization management, shareholder expectations, and regulatory oversight, thereby involving a broad range of stakeholders and considerations.

According to statistics, since the emergence of the “computing power leasing” concept in A-shares at the end of 2023, the number of companies under this concept has reached 93 as of mid-May this year, with more firms continuously announcing their entry into the sector.

So-called computing power leasing is currently concentrated primarily in the intelligent computing sector. Clearly, such large-scale investments by numerous enterprises not only boost stock prices but also deliver tangible economic benefits to these companies by enabling them to enter the intelligent computing industry.

In particular, unlike state-owned intelligent computing initiatives that place greater emphasis on social benefits, private-sector intelligent computing ventures prioritize tangible profits.

So, what are the revenue models for AI computing companies?
 

Unlike traditional data centers or cloud services, the core of the intelligent computing industry lies in delivering compute-as-a-service. Consequently, its business model combines the characteristics of data-center equipment leasing with the pay-as-you-go, usage-based pricing typical of cloud services. As a result, intelligent-computing companies’ business models exhibit a rich and diverse array of features.

Computing power leasing is one of the primary revenue models for intelligent computing centers. Instead of building their own costly computing infrastructure, enterprises and research institutions can rent computing resources from these centers to perform data processing and model training. This model lowers the barrier to entry, enabling more innovators to enter the AI field and accelerating the widespread adoption and development of AI applications.

Currently, most companies entering the intelligent computing sector also generate revenue through intelligent computing leasing. For example, Lianhua Zixing, a subsidiary of Lianhua Health, has been engaged in the computing power business since its entry into the field; in accordance with requirements from the Henan Securities Regulatory Bureau, it is required to disclose its monthly profitability. As of April alone, Lianhua Zixing entered into “Computing Power Leasing Service Contracts” with Company V and Company W, and during the same month collected a total of RMB 5.7686 million in payments for previously provided computing power services to Companies K, O, S, and V.

If computing power leasing is merely a resource service with no particular distinguishing features, then AI large-model training and deployment services are precisely what define an intelligent computing center and represent the original vision behind building such infrastructure. Intelligent computing centers can provide enterprises with a comprehensive suite of related services and technical support. By leveraging the high-performance computing resources of these centers, companies can rapidly train and optimize AI models and subsequently deploy them in real-world application scenarios.

Currently, major cloud service providers are the primary suppliers of large-model training and deployment services. Tencent Cloud MaaS, Baidu Smart Cloud, Huawei AI Cloud Services, and China Telecom’s computing-power service ecosystem platform, among others, all offer these services to customers. In addition, UFIDA has launched an enterprise-service-specific large model. YonGPT Specifically, it leverages China Telecom’s full-stack computing services, encompassing computing resources, computing networks, and computing platforms, while also supporting the optimization and enhancement of UFIDA’s large language model.

Large-scale models are undoubtedly the hottest application in today’s intelligent computing centers. In reality, however, various AI services have gradually permeated every aspect of daily life in recent years—ranging from facial recognition and document recognition to autonomous driving, smart healthcare, intelligent manufacturing, smart cities, and smart financial services—and these applications have already evolved into mature markets.

Computing power, algorithms, and data are the three core components of artificial intelligence. Many AI computing companies inherently possess vast data resources; under the premises of legality, security, and privacy protection, they can provide enterprise clients with services such as large-model datasets, data annotation, pre-training, and fine-tuning. In general, these services are particularly critical for data-intensive industries, including finance, healthcare, and retail.

The business model of intelligent computing centers is continuously evolving alongside advances in AI technology. Ranging from compute power leasing and data processing to the training and deployment of AI models, these centers are becoming indispensable infrastructure in the digital age. As intelligent computing centers are further refined and optimized, they will provide even stronger support for society’s drive toward intelligent transformation.

With the advancement and widespread adoption of AI technologies, the business models of the intelligent computing industry will continue to evolve. Ranging from computing power leasing and data processing to the training and deployment of AI models, intelligent computing is becoming an indispensable infrastructure in the digital era. Building a comprehensive and efficient intelligent computing ecosystem will provide even stronger support for China’s smart transformation of its economy and society.