AI Data Centers: A New Era for Power

AI data centers are rapidly evolving, with an increasing focus on energy efficiency and the ability to handle the high power demands of modern GPUs. An emerging trend is the adoption of 800V DC (direct current) systems and solid-state transformers.

This transition is driven by the need to reduce energy losses during the conversion of alternating current (AC) to direct current (DC), an inevitable process for powering servers and GPUs. 800V DC systems allow for the reduction of currents and, consequently, Joule effect losses in cables and components.

Solid-state transformers (SSTs) represent a more efficient and compact alternative to traditional low-frequency transformers. SSTs use power electronic components to convert voltage, offering greater flexibility and control, as well as a reduction in size and weight.

The adoption of these technologies is fundamental to support the growing power demand of AI workloads, while ensuring greater energy efficiency and a reduction in environmental impact. For those evaluating on-premise deployments, there are trade-offs to consider carefully, as discussed in AI-RADAR's analytical frameworks on /llm-onpremise.