The computing power industry is entering an AI supercycle driven by the scaling law of large models, the scaling of inference applications, and the deployment of intelligent agents. Unlike the pulsed demand of the previous cloud computing cycle, the current demand for computing power has shifted towards long-term, massive, and rigid growth. Global tech giants and leading cloud providers are resuming capital expenditure expansion, with data centers, as AI infrastructure, becoming a core beneficiary. This qualitative change in demand is rewriting the rules of the data center industry. In the traditional IDC era, competition centered on cabinets, location, and customer resources; in the AI era, the core of competition has evolved to encompass power acquisition capability, high-density construction capability, financial strength, engineering delivery efficiency, and long-term operational stability. The restructuring of the industry's underlying logic inevitably leads to a transformation in valuation frameworks. Data centers should no longer be viewed as static, mature, low-growth "rental-type" utility assets but should be defined as high-growth infrastructure platforms at the convergence point of three cycles: capital expenditure, technological upgrades, and demand expansion. Their value should not be assessed solely on single-period EBITDA or current utilization rates; instead, a dynamic multi-stage model should be applied: focusing on orders and delivery in the short term, utilization ramp-up and EBITDA expansion in the medium term, and free cash flow, return on capital, and asset recycling capability in the long term. The entity that can consistently convert AI computing power demand into new capacity, revenue growth, and cash flow expansion will achieve long-term value reassessment. As a leading domestic data center platform, GDS-SW (09698), with its four core barriers—financial reserves, customer orders, resource layout, and delivery operations—stands as one of the highest-certainty beneficiaries in this AI infrastructure boom cycle.
Financial strength builds competitive barriers, with diversified financing supporting cyclical expansion. Data centers are a capital-intensive industry, requiring significant investment in power supply, cooling, and engineering, with large single-project investments and long construction cycles. Capital strength has become a core entry barrier. As of the end of the first quarter of 2026, GDS-SW held cash and cash equivalents of RMB 14.823 billion. The company has established a diversified financing system covering equity, debt, and asset securitization both domestically and internationally. In the first quarter, the company realized USD 385 million by selling a portion of its equity in financial investment target DayOne Data Centers, retaining a 19.9% stake post-transaction, corresponding to a latest Series C valuation exceeding USD 2.2 billion. This transaction realized investment returns, bolstered cash flow, and retained growth benefits from overseas assets. The company allocated all realized funds to the AI infrastructure sector, focusing on its core business and optimizing capital allocation. Additionally, the company completed ABS asset securitization projects and its first C-REIT issuance in 2025, and in early 2026, raised USD 300 million by issuing convertible preferred shares to Huatai Capital, further enhancing its financial reserves. Ample cash reserves, diverse financing channels, and mature asset monetization capabilities provide GDS-SW with significant first-mover advantages and anti-cyclical resilience during industry expansion windows. When computing power demand surges, small and medium-sized IDC enterprises are often constrained by capital scale, financing costs, and balance sheet pressure, struggling to expand production and deliver promptly; in contrast, GDS-SW possesses the capability to rapidly accommodate major clients' computing power demands, advance large-scale project construction, and continuously optimize its capital structure.
Order fulfillment validates the rigidity of demand, with deep collaboration from leading clients. AI computing power demand has fully transitioned from an industry trend to tangible, actionable orders. In the first quarter, GDS-SW added a record-high 200 MW of net new signings. During the same period, the company's signed and pre-signed area increased by 11.7% year-over-year, utilized area grew by 12.7% year-over-year, and operational area utilization rate rose to 77.3%. As of the end of the first quarter, the company's operational area commitment rate remained at a high level of 92.8%, and the pre-commitment rate for area under construction reached 84.4%. To date, the company's total signed scale for the year has approached 350 MW, surpassing the total signed volume for the entire year of 2025. For data center enterprises, signings, pre-signings, and utilization rates are the most critical leading indicators. Leading clients locking in large-scale computing power resources in advance signals a recovery in long-term computing power demand; the steady increase in utilized area and utilization rates indicates that existing assets are being converted into stable revenue.
Capital-Intensive + Strong Operations: GDS-SW Aggressively Advances AI Infrastructure. Order landing is only the first step; stable delivery, continuous operation, and long-term monetization constitute the long-term core competitiveness of data center enterprises. The core logic is that data centers are no longer about stacking cabinets but involve comprehensive competition in high-power density, stable power supply, engineering efficiency, and long-term operational maintenance capabilities. In core regions, high-quality land, power, and compliance resources are becoming increasingly scarce, enhancing the value of platforms with resource acquisition and large-scale delivery capabilities. To address the stratified computing power demands of clients, GDS-SW has built an efficient infrastructure network through coordinated layout in "mature markets + emerging hubs." In mature markets around the Beijing-Tianjin-Hebei region, Yangtze River Delta, Guangdong-Hong Kong-Macao Greater Bay Area, and Southwest China, the company leverages low latency and high network quality advantages to secure core computing power demands with stringent real-time requirements. Simultaneously, in emerging hub nodes such as Ulanqab, Hohhot, and Zhongwei, it utilizes energy and cost advantages to deploy large-scale heterogeneous computing power clusters for training and inference. GDS-SW has established a complete closed loop from resource acquisition and project construction to customer signing, capacity delivery, and long-term operation. As of the end of the first quarter, the company's operational area reached 674,300 square meters, a year-over-year increase of 10.4%; area under construction reached 118,400 square meters, a quarter-over-quarter increase of 60.0%. To accommodate incremental AI computing power demand, the company announced plans to invest RMB 30 to 50 billion in capital expenditure over the next three years for data center construction. This scale of capital expenditure not only sets a new historical high for the company but also fully reflects confidence in the long-term demand for AI computing power.
The Long AI Infrastructure Cycle Begins, Establishing a "The Strong Get Stronger" Landscape. In the coming years, the data center industry will enter an expansion phase, with industry growth dividends increasingly concentrating towards leading platforms possessing comprehensive competitive advantages. The survival space for small and medium-sized players will further narrow, a trend that will amplify GDS-SW's core competitiveness. GDS-SW has fortified its financial moat through ample cash reserves and a diversified financing system, supported by strong order growth from leading clients, solidified its production capacity base with high-quality resource reserves in core regions, and ensured efficient delivery and stable operations through its full-chain closed-loop operations, aligning perfectly with the core competitive logic of the data center industry in the AI era. As the rigid demand for AI infrastructure is unleashed, the company is poised to fully capture industry dividends with its comprehensive platform capabilities, achieving steady expansion in capacity, revenue, and cash flow, and ushering in a Davis double-click on both earnings and valuation.
Comments