Danish pump and water solutions company Grundfos has identified China as a critical market for advanced data center cooling solutions amid the rapid growth of artificial intelligence (AI) technologies. The company’s vice president for industry solutions and marketing, Julian Trascasa Cano, highlighted that the rise in computing density driven by AI is increasing demands on electricity and water resources, presenting challenges that require innovative cooling approaches.

Data center solutions have evolved from a specialized niche to one of Grundfos’ principal business areas in recent years. Cano anticipates robust demand growth over the next three years, with China representing a significant portion of this expansion. He pointed to the exponential rise in global data needs, which is accelerating with developments in AI and digital twin technologies, creating what he calls a “density dilemma” related to managing power and cooling in increasingly compact computing environments.

China’s computing capacity has surged, reaching 2,185 exaflops by the end of June 2026—a 177% increase compared to the previous year—according to the Ministry of Industry and Information Technology. The country’s computing facilities operated at an average utilization rate of 71.4%, underscoring the scale of data center activity.

The rapid growth is prompting a shift in the definition of green data centers in China. The China Academy of Information and Communications Technology (CAICT), in collaboration with Grundfos, released a joint white paper during an industry conference in Beijing that addresses the evolving requirements. The paper emphasizes that future data center sustainability depends on integrated management of computing power, energy, and water resources, rather than focusing solely on energy efficiency.

Traditional air cooling methods are being outpaced as AI chips consume more power and data racks become denser. Nearly all electricity used by data centers converts to heat, making fluid-based cooling systems increasingly vital for operational efficiency and equipment safety, Cano explained. Liquid cooling methods, while more effective at heat transfer than air cooling, do not inherently reduce overall energy or water consumption. Their environmental impact will depend on design choices such as cooling system architecture, local climate, water availability, and the responsiveness of pumps and equipment to changing computational loads.

This issue is particularly acute in China’s western regions, where cooler climates and renewable energy availability can lower cooling energy demands, but water scarcity poses challenges for sustainable operation. The white paper proposes a three-layer framework covering system architecture, cooling equipment, and digital management platforms. It advocates for dynamic cooling distribution based on demand, intelligent pump controls, leak detection, heat recovery, and closed-loop system adjustments.

Cano noted that enhanced coordination among computing loads, cooling infrastructure, and water management could reduce overall electricity consumption by up to 12% in well-operated data centers without significant hardware upgrades, though outcomes would vary by facility.

Grundfos’ focus on data centers has supported its financial growth, with the company reporting revenue of 17.9 billion Danish krone ($2.78 billion) in the first half of 2026, a 7.7% increase in local currency sales. Data center solutions were a major growth driver, especially in the United States, where sales increased by more than 20%.

In China, the company has established integrated operations including product development, manufacturing, and delivery with factories in Jiangsu province and Shanghai. Grundfos is also collaborating with Chinese clients to develop digital pump platforms and advanced control systems that adjust cooling output in real time. Some of these locally developed technologies are now being introduced in other Asian and Western markets.

While conventional data centers still consume more energy and water overall, the AI workload segment is growing at a faster pace and is expected to become the largest driver of additional energy and cooling needs. Cano emphasized that meeting this challenge requires responsible expansion of computing capacity and collaborative solutions across the industry.