In the context of deepening "AI+" initiatives and accelerating domestic substitution, Dawning Information Industry Co., Ltd. (SH603019), a leading A-share computing power company, has unveiled a significant refinancing plan. On the evening of February 9, the company announced that its board of directors approved a plan to issue convertible bonds to unspecified investors, aiming to raise up to 8 billion yuan.
This marks a rare large-scale fundraising effort by Dawning Information since its listing. The proceeds will be entirely allocated to advanced AI computing clusters, integrated training-inference systems, and domestically developed advanced storage systems.
However, amid intense competition from industry giants and rapid technological advancements, this debt-financed expansion represents both a major bet on future computing demand and a strategic move to navigate risks of earnings dilution and fierce market rivalry.
The 8 billion yuan fundraising will target three key projects: "Advanced Computing Cluster Systems for Artificial Intelligence," "Next-Generation High-Performance AI Integrated Training-Inference Systems," and "Domestically Developed Advanced Storage Systems." These interconnected initiatives outline the company's ambition to build an integrated ecosystem spanning computing power, algorithms, and storage.
According to the proposal, the largest portion of the funds—3.5 billion yuan—will be directed toward the advanced AI computing cluster project. This initiative goes beyond simple hardware aggregation, focusing on developing ultra-node hardware systems, high-speed interconnects, foundational software stacks, and heterogeneous computing resource management platforms based on high-performance computing architectures and tight hardware-software integration.
The surging demand for AI computing power provides substantial market potential for these projects. IDC data indicates that the global AI server market is projected to grow from approximately $125.1 billion in 2024 to $158.7 billion in 2025, potentially reaching $222.7 billion by 2028.
As AI models expand from millions to trillions of parameters, explosive growth in computing demand has led to severe global shortages and resource allocation imbalances. Dawning Information believes that establishing efficient, coordinated computing resource scheduling capabilities is essential to address fragmentation, improve utilization, tackle energy challenges, and meet diverse application needs.
The second project, involving 2.5 billion yuan, focuses on next-generation integrated AI training-inference systems. To overcome key challenges in AI deployment—high technical barriers, unpredictable costs, and lengthy cycles—the company aims to deliver "out-of-the-box" solutions through integrated hardware-software co-design. By developing a full-stack approach encompassing chips, algorithms, and hardware, the project seeks to bridge domestic AI hardware with application layers, forming an industrial loop from chips to end-user scenarios and enhancing AI computing efficiency.
Additionally, 2 billion yuan is earmarked for developing advanced domestic storage systems. The company stated that this project will fully unleash computing chip performance, handle massive multimodal data processing requirements, and provide foundational storage capabilities essential for AI applications.
Dawning Information emphasized that the fundraising projects align with its core business, national industrial policies, and industry trends. They reflect the company's technology-driven development strategy and are expected to yield favorable market prospects and returns.
Despite the promising AI sector, the substantial fundraising highlights the delicate balance between growth objectives and financial needs. Financial disclosures show that Dawning Information maintained consecutive annual net profit growth from 2022 to 2024, with net profits attributable to shareholders of 1.544 billion yuan, 1.836 billion yuan, and 1.911 billion yuan, respectively. However, 2024 revenue declined to 13.148 billion yuan, down 8.40% from 14.353 billion yuan in 2023, which the company attributed to intensified competition and strategic reduction in low-margin projects.
Of particular note is the company's cash flow situation. For the first nine months of 2025, net cash flow from operating activities was negative 1.296 billion yuan. While the company explained this is typical due to higher customer repayments in the fourth quarter, the 8-billion-yuan capital expenditure plan will further test its financial management capabilities.
Dawning Information acknowledged that expanding R&D and business scale increases working capital needs, requiring sufficient funding to support operations and enhance competitiveness.
The company previously raised 4.78 billion yuan through a 2020 private placement, which was fully utilized by the end of 2023 and achieved expected returns. However, the potential dilutive effect of the new convertible bond issuance remains a concern for investors. The company cautioned that the conversion of bonds might lead to a short-term decline in metrics like earnings per share if project benefits are not realized promptly.
To mitigate dilution risks, Dawning Information committed to strengthening fund management and maintaining consistent dividend policies. Over the past three years (2022–2024), the company distributed 1.113 billion yuan in cash dividends, representing 63.09% of average distributable profits. It plans to continue adhering to its dividend policy post-issuance.
Notably, the feasibility report repeatedly emphasized "independent controllability" and "domestic substitution." With global tech giants accelerating布局 in integrated AI systems, domestic firms currently compete mainly via "general servers + GPUs," lacking deeply adapted scenarios and integrated ecosystems. Whether Dawning Information can establish a technological moat spanning chips, algorithms, and hardware through these projects will be critical to the 8-billion-yuan investment's returns.
Comments