Factors such as extreme densification, gigawatt-level scalability and the data center as a computing unit are driving these trends
Data center innovation continues to be shaped by macroeconomic scenarios and technology trends related to artificial intelligence, according to a report from Vertiv (NYSE: VRT), a global leader in critical digital infrastructure. The report Vertiv™ Frontiers, which draws on expertise across the organization, details the technology trends driving current and future innovation, from power for AI to digital twins to adaptive liquid cooling.
“The industry continues to rapidly evolve in data center design, construction, management and support in response to the density and speed of AI factory deployment demands,” said Scott Armul, chief product and technology officer at Vertiv. "We see forces across technologies, including extreme densification, driving transformative trends such as high-voltage direct current (DC) power architectures and advanced liquid cooling systems, critical to achieving gigawatt-level scalability and critical to AI innovation. On-site power generation and digital twins should also help accelerate the scale and speed of AI adoption."
The Vertiv Frontiers report builds on and extends the Vertiv's previous annual data center trends forecast. The document identifies the macro-factors driving innovation: extreme densification — accelerated by AI and HPC workloads; scalability to gigawatts at high speed — data centers are now being deployed rapidly and at unprecedented scale; data centers as computing units — the AI era requires structures designed and managed as a single system; and silicon diversification — data center infrastructure must adapt to a growing range of chips and computing solutions.
The report illustrates how these macro-trends have in turn influenced five key trends impacting specific areas of the data center industry.
- Enhanced power for AI
Most current data centers still rely on hybrid AC/DC power distribution from the network to the IT racks, which includes three or four conversion stages and some inefficiencies. This approach is under pressure as power densities increase, largely driven by AI workloads. Moving to higher voltage DC architectures enables significant reductions in current, conductor size and number of conversion stages, centralizing energy conversion at the room level. Hybrid AC and DC systems are popular, but as all-DC standards and equipment mature, high-voltage DC is set to become more common as rack densities increase. On-site generation and microgrids will also drive the adoption of high-voltage DC.
- AI distributed
The billions of dollars invested so far in AI data centers to support large language models (LLMs) have been intended to support the widespread adoption of AI tools by consumers and businesses. Vertiv believes that AI is becoming increasingly critical to businesses, but how and where such inference services are delivered will depend on the specific requirements and conditions of the organization. This will impact businesses of all types; however, highly regulated industries such as finance, defense, and healthcare may need to maintain private or hybrid AI environments via on-premise data centers, due to data residency, security, or latency requirements. Flexible, scalable high-density power and liquid-cooling systems can enable capacity through both new construction and retrofits of existing facilities.
- Acceleration of energy autonomy
Short-term on-site power generation capacity has been essential to the resiliency of most standalone data centers for decades. However, widespread power availability challenges are creating the conditions to adopt extended power autonomy, especially for AI data centers. Investments in on-site generation, via natural gas turbines and other technologies, have several intrinsic benefits but are mainly driven by the critical issues of energy availability. Technological strategies such as Bring Your Own Power (and Cooling) are destined to become part of energy autonomy plans.
- Digital twin-driven design and operations
Increasingly dense AI workloads and more powerful GPUs require the rapid deployment of complex AI factories. Using AI-based tools, data centers can be virtually mapped and specified via digital twins; IT and critical digital infrastructure can be integrated, often as prefabricated modular designs, and distributed as computing units, reducing the time-to-token up to 50%. This approach will be critical to efficiently realizing the gigawatt-scale expansions necessary for future AI advances.
- Adaptive and resilient liquid cooling
AI workloads and infrastructures have accelerated the adoption of liquid cooling. At the same time, AI can be used to further refine and optimize these solutions. Liquid cooling has become mission-critical for a growing number of operators, and AI may offer ways to further enhance its capabilities. Combined with additional monitoring and control systems, AI has the potential to make liquid cooling systems smarter and more robust, predicting potential failures and effectively managing fluids and components. This trend should lead to greater reliability and availability (uptime) for high-value hardware and associated data/workload.
Vertiv operates in more than 130 countries, providing critical digital infrastructure solutions to data centers, communications networks, and commercial and industrial facilities around the world. The company's comprehensive portfolio includes solutions and services for energy management, thermal management and IT infrastructure, from the cloud to the network edge. This integrated approach enables continuous operations, optimal performance and scalable growth for customers facing an increasingly complex digital landscape.
