AI growth reshapes data center infrastructure and energy needs
The expansion of artificial intelligence is increasing the energy demands of data centers and accelerating changes in their design.
The growth of AI workloads is driving a transformation of digital infrastructure, creating new demands for power, cooling and computing capacity.
The advancement of artificial intelligence is transforming the design and operation of data centers, which now face greater demands for power, cooling and processing capacity to support new workloads.
For years, these facilities were primarily focused on storing and processing data. However, the expansion of generative models, advanced automation and real-time data analytics is changing the requirements of digital infrastructure.
The growth of artificial intelligence applications is driving higher energy demand and forcing the redesign of data center infrastructure.
Rising energy demand.
The projections illustrate the scale of this transformation. A 2025 report by the International Energy Agency (IEA) estimates that global electricity consumption by data centers could more than double by 2030, reaching nearly 945 TWh, a level equivalent to Japan’s annual electricity consumption.
Meanwhile, McKinsey & Company projects that global demand for data center capacity will nearly triple by the end of the decade and that around 70% of that capacity will be dedicated to artificial intelligence workloads.
This scenario is shifting the industry’s focus from computing capacity to power availability, energy efficiency, grid stability and cooling strategies.
These challenges were discussed during AI Ready DC 2026, an event organized by Schneider Electric to examine the impact of artificial intelligence on critical infrastructure.
Luis Santamaría, Cloud & Service Provider Lead at Schneider Electric, explained:
”Today, the conversation is already shifting toward infrastructure designed specifically for AI rather than traditional data centers. This brings enormous electrical, energy and engineering challenges. We are already seeing 150-kilowatt racks, and projections point to as much as one megawatt per rack in the coming years, representing a major challenge for electrical infrastructure and cooling.”
New cooling models.
Higher power densities are also driving changes in cooling systems. According to Santamaría, technologies such as liquid cooling are beginning to emerge as an alternative for meeting new thermal requirements.
Santamaría added:
”Dissipating and reusing that amount of heat requires new design strategies, greater engineering expertise and solutions such as liquid cooling, because traditional air-based systems are no longer sufficient.”
The executive also said that artificial intelligence is reshaping data center architecture through servers with greater computing capacity and more compact data halls, while requiring increasingly higher levels of power and cooling.
In practice, the expansion of AI is driving the evolution of data centers, where energy and cooling management are becoming just as important as computing capacity.
Infrastructure for AI.
In this context, Schneider Electric said it has strengthened its collaboration with NVIDIA to develop infrastructure architectures for artificial intelligence workloads and integrate power, cooling, electrical distribution and digital management solutions for high-performance computing environments.
The company said that developing AI-ready data centers also requires an ecosystem of specialized partners capable of addressing all aspects of these facilities, particularly in markets such as Colombia and Chile, where the expansion of renewable energy and continued technological progress are supporting the development of next-generation digital infrastructure.
|
|
