AI-Driven Expansion of Megawatt-Scale Data Centers Puts Energy Efficiency in Global Spotlight

As AI models grow in scale and computing demand keeps rising, the world has entered a new phase of megawatt-scale data center buildout. According to the International Energy Agency (IEA)^{1 2}, global electricity consumption by data centers is expected to double within five years, reaching 945 TWh (terawatt-hours), equivalent to the current total annual power consumption of Japan. 
In this context, improving data center energy efficiency (measured by PUE, Power Usage Effectiveness) has become a shared goal across the industry. Countries around the world are actively seeking technical solutions that can reduce power distribution losses and improve conversion efficiency. The broader trend is clear: data center power architectures are rapidly moving toward higher-voltage, higher-efficiency designs, with power transmission efficiency becoming a central issue in AI infrastructure development.

HVDC Becomes Mainstream, SSTs Play Key Role in Power Delivery

One effective way to lower PUE is to reduce the amount of energy lost as heat during power transmission and conversion. Under Joule’s law, this type of transmission loss increases with the square of the current. For a given power level, raising the voltage reduces the current required, in turn improving efficiency. This is why data center power architectures have progressed from 12V (ORv1) through 48V (ORv3) to today’s HVDC ±400V and 800V. 
As the industry moves to adopt these DC voltage solutions, solid-state transformers (SSTs) have attracted growing attention in major AI forums such as the Open Compute Project (OCP). As shown in the figure below, SSTs can support high-voltage AC input from 10 to 35+ kVAC and deliver outputs such as ±400V, 800V, or other application-specific voltages.³ This architecture is well-suited to AI data centers, renewable grid integration, and next-generation power conversion systems, and is set to play an important role in the next wave of energy infrastructure.

Chroma Offers Comprehensive SST Module Test Portfolio Covering AC and DC Requirements

A solid-state transformer consists of multiple integrated power modules. By combining these modules in series and parallel, the system can handle medium-voltage AC (MVAC) inputs and ±400V/800V outputs, providing megawatt-level power capacity. To help customers shorten development cycles and enhance the reliability of SST modules, Chroma has introduced a comprehensive solution covering AC-side simulation and DC-side load testing.

• AC-side test solution
  The Chroma 61800-100 Grid Simulator, paired with a dedicated step-up transformer, delivers up to 1600 VLL single-phase output, meeting the voltage requirements of most module-level tests and enabling customers to fully simulate and validate a wide range of grid conditions.
• DC-side test solution
  The Chroma 62450D-2000HL Programmable Bidirectional DC Power Supply supports both 2000VDC source output and regenerative load operation. This makes it suitable for the voltage requirements of most DC links in SST modules as well as bidirectional architecture testing, while reducing test energy consumption by more than 80%.
  The Chroma 63224A-1200 High Power Programmable DC Electronic Load provides 1200V capability together with high-current microsecond-level slew rate performance, allowing users to carry out the fast dynamic response load testing required by most devices under test.

Key SST Testing Capabilities for a Stable, High-Efficiency AI Power Infrastructure

With more than 40 years of experience in automatic test equipment (ATE) and precision measurement instruments, Chroma continues to support customers across every stage of SST module development. From R&D and design verification to mass production, our engineers stand ready to help you achieve outstanding quality and reliability. Get in touch with our team or click the links below for more information on relevant products and applications.

Sources and Notes:
[1] Lin Boyu, report published April 11, 2025.
https://www.reccessary.com/en/news/iea-energy-and-ai-report 
[2] International Energy Agency (IEA), published April 11, 2025.
https://www.iea.org/news/ai-is-set-to-drive-surging-electricity-demand-from-data-centres-while-offering-the-potential-to-transform-how-the-energy-sector-works
[3] Image source: “2025 OCP APAC Summit,” by Peter Barbosa, Delta Electronics. Original materials provided by OpenCompute.org. Licensed under CC BY-SA 4.0. Adapted by Chroma ATE Inc.