GDS-SW(09698): AI-driven business accelerates expansion, with a scale of billions invested in domestic data center construction.

On August 13, GDS-SW (09698), a leading developer and operator of high-performance data centers in China, announced its unaudited financial results for the second quarter ending June 30, 2026. The financial report shows that the company achieved a net revenue of 3.088 billion yuan (RMB, the same below) in the second quarter, a year-on-year increase of 6.5%. Due to the rising proportion of utility costs, the gross margin decreased by 2.3 percentage points year-on-year to 21.5%. The net profit attributable to the parent company was 838 million yuan, with a net profit margin of 27.1%. The adjusted EBITDA was 1.406 billion yuan, a year-on-year increase of 2.5%, with an adjusted EBITDA margin of 45.5%, a decrease of 1.8 percentage points compared to the previous year. The market responded positively to the company's operational performance in the second quarter. On August 14, GDS shares opened sharply higher on the Hong Kong stock exchange, reaching a peak of 35.16 Hong Kong dollars during the trading session, marking a maximum increase of 13.3%. AI demand becomes a growth engine, annual performance guidance raised It has been observed that the market's optimistic sentiment partially stems from the expectations created by GDSs raised performance guidance. After adjustment, the company estimates that its net revenue for the full year of 2026 will be in the range of 12.7 billion to 13 billion yuan, an increase of 11.1% to 13.7% year-on-year; the adjusted EBITDA is projected to be in the range of 5.9 billion to 6.1 billion yuan, an increase of 9.2% to 12.9% year-on-year. Several core operational metrics showed impressive performance in the first half of the year, providing support for growth in the second half. As of June 30, 2026, the total signed and pre-signed area was 784,800 square meters, a year-on-year increase of 18.2%; the billable area was 542,200 square meters, an increase of 13.2% year-on-year; the current billing rate was 79.2%, an increase of 1.7 percentage points year-on-year. At the end of the second quarter, the pre-signing rate for construction projects reached 89.2%, an increase of 14.5 percentage points year-on-year. According to the characteristics of the industry, the signed and pre-signed area and billing rates are leading indicators for observing the medium- and long-term performance of data center enterprises. Leading cloud service providers and large internet companies usually lock in large-capacity data center resources several years in advance, and the increase in signing scale directly reflects downstream computing power demand; a rising billing rate indicates that signed cabinets are gradually completing server deployments, and previous contract orders are transitioning into operating cash flow. In terms of incremental orders, the companys management noted in the second quarter earnings call that the company signed a new computing capacity of 260 MW in the second quarter of 2026, with a total of 470 MW signed in the first half of the year, setting a historical record. As of the end of the second quarter, the company has secured an intention to reserve 600 MW of computing power resources, with enforceable orders totaling 757 MW, predominantly on a take-or-pay basis. In light of the strong sales momentum driven by AI demand, the company has raised its annual sales target to 1 GW. GDS Chairman and CEO Huang Wei stated that in the second quarter, while advancing order delivery, the company maintained a high level of new signed orders. Based on current progress, the total sales signed by the company this year is expected to set a record, far exceeding the initially set target. The company is very optimistic about the market opportunities brought by domestic AI demand and firmly believes in its ability to seize this opportunity to drive large-scale business growth. Optimizing capital structure, investing 30 to 50 billion yuan in data center construction With the demand for business expansion, GDS plans to invest 30 to 50 billion yuan in data center construction over the next three years. The company has raised its capital expenditure guidance for 2026 from approximately 9 billion yuan to 10 billion yuan. The data center industry is capital-intensive, with long project construction cycles and significant financial investments; increasing capital expenditures will test the company's ability to raise funds and manage capital allocation. In recent years, GDS has established a diversified financing system that includes equity, debt, and asset securitization, supplementing funds and optimizing capital allocation through various channels. In 2025, GDS completed an ABS asset securitization project and the first issuance of C-REITs, unlocking asset securitization financing channels; and in early 2026, it issued $300 million of convertible preferred shares to Huatai Capital, supplementing long-term funds. As of the end of the second quarter of 2026, the company's cash and cash equivalents reached 14.927 billion yuan, a year-on-year increase of 13.7%; at the same time, the company's debt-to-asset ratio was 61.25%, a year-on-year decrease of 6.44 percentage points, indicating improved financial stability. Additionally, through an external investment, GDS recognized an investment income of 960 million yuan in the second quarter using the equity method. If the related assets realize fund flows in the future, it will support the construction of domestic data center projects. The diversified financing channels and mature asset monetization capabilities help GDS gain a first-mover advantage during the industry's expansion window and enhance its resilience against economic cycles. During the concentrated release phase of computing power demand, small and medium IDC enterprises may be constrained by funding scale and financing costs, making it difficult to expand production and deliver on time; in contrast, leading IDC companies possess the comprehensive ability to quickly meet large customer computing power needs, advance project construction rapidly, and continuously optimize their capital structure. Targeting AI computing power demand, strategically positioning resource layouts In terms of resource layout, GDS adopts a "mature market + emerging hub" model to cater to the differentiated computing power demands of customers. Data centers in core economic circles such as Beijing-Tianjin-Hebei, the Yangtze River Delta, the Greater Bay Area, and Chengdu-Chongqing primarily support businesses with high latency requirements, such as financial transactions and real-time inference; while national-level hub nodes like Ulanqab, Zhongwei, and Hohhot focus on large-scale computing clusters for offline training of large models. These two types of business scenarios complement each other, releasing synergistic value across regions. In June 2026, the company signed a strategic cooperation agreement with the Ulanqab municipal government to invest over 30 billion yuan over the next five years to create a large-scale GW-level zero-carbon intelligent computing cluster. The significance of positioning a computing hub in the west lies in obtaining scarce large-capacity electricity quotas while leveraging local energy advantages to reduce long-term operational costs. Currently, leading cloud service providers have incorporated renewable computing power into their procurement standards, and a high proportion of green electricity helps enhance the companys competitiveness in securing ultra-large-scale orders. Coupled with local energy consumption and power grid matching policy support, early positioning of western hubs can help the company reserve stable long-term capacity and optimize the electricity cost structure of AI operations. GDSs business layout reflects the changing competitive logic of the data center industry. Industry competition is no longer merely a matter of cabinet scale and geographical location; the ability to acquire electricity resources, the construction level of high-density data centers, financial reserves, rapid project delivery capability, and fine asset operation capabilities have become core competitive barriers, driving a reshaping of the overall valuation system in the industry. This round of expansion in the computing power infrastructure industry is driven by the iteration of large models, the comprehensive penetration of inference scenarios, and the implementation of multi-agent applications, presenting a long-term rigidity in demand that significantly differs from the short-term pulse-like trends seen in the past cloud computing phase. Global technology companies continue to increase capital investments in computing power infrastructure, with data centers, as the AI infrastructure, becoming the core beneficiaries.