VNET Group, a leading carrier-neutral data center services provider in China, has reported strong financial results for the first quarter of 2026, highlighting a significant structural shift in its business model driven by the artificial intelligence boom. According to the company’s official earnings release, total net revenues increased by 19.8% year-over-year to RMB 2.69 billion.
The most notable milestone in the Q1 report is that revenue from VNET’s wholesale data center business surpassed its retail colocation revenue for the first time in the company’s history. This crossover underscores the massive scale of infrastructure deployment currently underway in China to support the training and operation of large language models and other advanced AI applications.
The Shift to Wholesale
Historically, VNET’s business was heavily weighted toward retail colocation, providing server space, power, and connectivity to a diverse base of enterprise customers. However, the rise of cloud computing and, more recently, the explosive growth of generative AI, have fundamentally altered the demand profile for data center services.
Major tech giants, cloud service providers, and well-funded AI start-ups require massive, contiguous blocks of computing capacity. These clients prefer wholesale arrangements, leasing entire data halls or even dedicated facilities, enabling them to deploy thousands of high-density server racks optimized for AI workloads. VNET’s strategic pivot to cater to this wholesale demand is now clearly reflected in its financial performance, validating its investments in large-scale, high-power-density campuses.
Capacity Expansion and Utilization
To meet this surging demand, VNET has been aggressively expanding its footprint. The company reported that its total capacity in service reached 907 megawatts (MW) by the end of the first quarter, a significant increase from the previous year. Crucially, the utilization rate for this expanded capacity remained high, indicating that new supply is being rapidly absorbed by the market.
The high utilization rates are particularly impressive given the specialized requirements of AI data centers. Training large models requires specialized cooling infrastructure, such as liquid cooling, to manage the immense heat generated by densely packed clusters of GPUs or specialized AI accelerators, such as Huawei’s Ascend chips. VNET’s ability to deliver and monetize these advanced facilities demonstrates its technical capabilities and strong relationships with key players in the Chinese tech ecosystem.
The Broader Infrastructure Landscape
VNET’s Q1 results provide a clear window into the physical reality of the AI revolution in China. While software algorithms and model architectures often dominate the headlines, the foundation of this progress rests on concrete, steel, and massive amounts of electricity. The data center industry is the critical enabler of China’s AI ambitions, and companies like VNET are positioned as the primary beneficiaries of this infrastructure buildout.
As the US-China Tech War continues to restrict access to the most advanced foreign hardware, Chinese companies are increasingly relying on domestic data centers to maximize the efficiency of the chips they can acquire. This dynamic ensures that demand for high-quality, high-density data center capacity will remain robust in the foreseeable future, providing a strong tailwind for VNET and its peers in the infrastructure sector.
Power Constraints and the Path Forward
One of the most significant challenges facing VNET and the broader data center industry in China is securing adequate power supply. The energy requirements of AI data centers are immense, and the rapid pace of capacity expansion is straining local power grids in some regions. VNET has been actively working with local governments to secure long-term power purchase agreements and to site new facilities in areas with abundant renewable energy resources, such as Inner Mongolia and Guizhou, which offer both lower electricity costs and favorable climatic conditions for cooling.
Looking ahead, VNET’s management has signaled continued investment in expanding its wholesale capacity to meet the pipeline of demand from AI and cloud computing clients. The company’s ability to secure land, power, and permits in strategic locations will be the primary bottleneck on its growth trajectory. With 907 MW already in service and a robust development pipeline, VNET is well-positioned to remain a key infrastructure partner for China’s AI industry, and its Q1 2026 results offer a compelling financial case for the data center sector as a long-term investment theme alongside the broader AI infrastructure buildout.
VNET’s Competitive Position in a Crowded Market
VNET operates in a competitive landscape that includes domestic peers such as GDS Holdings, ChinaData, and Chindata Group, as well as the in-house data center operations of China’s major cloud providers: Alibaba Cloud, Tencent Cloud, and Huawei Cloud. The key differentiator for VNET is its carrier-neutral status, which allows it to serve multiple cloud providers and enterprise clients without the conflict of interest that arises when a data center operator is also a direct competitor in the cloud services market.
The Q1 2026 milestone of wholesale revenues surpassing retail for the first time is a structural shift that mirrors the evolution of the US data center industry, where hyperscale wholesale demand has been the dominant growth driver for years. For VNET, successfully executing this transition, while managing the capital intensity of large-scale campus development and the operational complexity of serving high-density AI workloads, will be the defining challenge of the next phase of its growth. The Q1 results suggest the company is navigating this transition effectively, and its performance will be closely watched as a proxy for the health of China’s broader AI infrastructure buildout.