Powering AI at Scale: A New Era for Data Centers
AI clusters are reshaping the data center landscape. With rack densities and energy demand reaching unprecedented levels, U.S. operators face increasing pressure to deliver more compute per square foot while meeting ambitious sustainability objectives. I anticipate steady progress this year as we move closer to solving these challenges.
Power architecture is a strategic priority. As AI training and inference transition from pilot projects to full-scale production, traditional AC designs are being stretched to their limits. As a result, DC architectures are emerging as a long-term strategy in power design. Industry leaders recognize this approach as the most practical path to achieving efficiency, density, and scalability—unlocking megawatt IT racks and gigawatt campuses while reducing copper usage and conversion losses.
Key advantages driving this shift include:
| • | Efficiency at Scale: Fewer conversion stages improve end-to-end efficiency—critical when megawatts of IT load translate into millions in energy costs. |
| • | Higher Power Density: 800 V DC supports megawatt-class racks without the copper burden and space constraints of low-voltage systems. |
| • | Native Renewable Integration: Batteries and solar operate natively in DC, making HVDC an ideal platform for sustainable, resilient power strategies. |
At ABB, we’re excited to partner with you on this journey. We’re investing in our people, portfolio, and production to help build the world’s most advanced data centers—together.
