Growing demand for artificial intelligence (AI) is changing how data centers are designed, built, and run, according to a new report from critical digital infrastructure company Vertiv. The Vertiv Frontiers report outlines technology and operational trends that the company expects to shape data centers in the coming years, driven largely by AI workloads that require more power, higher density, and faster deployment.
Data centers are under pressure to support AI systems that consume far more computing power than traditional applications. These systems often rely on high-performance chips that generate significant heat and require new approaches to power delivery and cooling.
“The data center industry is continuing to rapidly evolve how it designs, builds, operates, and services data centers, in response to the density and speed of deployment demands of AI factories,” said Scott Armul, chief product and technology officer of Vertiv.
He added that the push for denser computing environments is leading to changes in core infrastructure.
“Cross-technology forces, including extreme densification, are driving transformative trends such as higher voltage DC power architectures and advanced liquid cooling that are important to deliver the gigawatt scaling that is critical for AI innovation,” Armul said.
The report highlights several broad forces affecting the industry. These include the rapid growth of AI and high-performance computing, the need to build very large data centers quickly, and the growing mix of different chips used for computing. As a result, data centers are increasingly treated as a single, integrated system rather than separate parts.
One major trend is the shift in how power is delivered. Many data centers still use a mix of alternating current (AC) and direct current (DC), which involves multiple power conversions and energy loss. As AI workloads push power demand higher, Vertiv expects more operators to move toward higher voltage DC systems. These systems can reduce energy loss and support denser equipment. On-site power generation and microgrids may also become more common as operators face limits in grid capacity.
The report also noted the rise of distributed AI. While large cloud-based AI systems remain important, some industries, including finance, defense, and healthcare, may need private or hybrid AI setups. These organizations often require tighter control over data location, security, and response time, which can favor on-site data centers.
The report also saw that power shortages in some regions are pushing data center operators to invest more in their own energy sources, such as gas turbines, to ensure reliability.
Vertiv also sees digital twin technology playing a bigger role. By creating virtual models of data centers, operators can plan and deploy facilities faster and integrate equipment more efficiently. This approach can cut deployment time significantly.
Liquid cooling is becoming essential for AI hardware. The report notes that AI tools themselves may help monitor and manage cooling systems, improving reliability and reducing downtime as computing demands continue to grow.