SolarEdge And Infineon Collaborate On AI Datacenter Transformer

This solid-state transformer targets the 800 V DC power architecture of hyperscale AI data centers
SolarEdge's SST technology.
The upcoming SST will be based on SolarEdge’s DC-coupled power conversion architecture. (Photo Credit: SolarEdge)
Published on
Key Takeaways
  • SolarEdge and Infineon are collaborating to develop a solid-state transformer that will directly convert medium voltage AC input to DC, eliminating multiple intermediate steps of the AC power architecture-based power distribution 

  • This transformer will use Infineon’s SiC-based switching devices 

  • It is expected to achieve a power conversion efficiency of more than 99% 

Power conversion solutions provider SolarEdge Technologies and semiconductor solutions maker Infineon Technologies are collaborating on the development of a solid-state transformer (SST) platform specifically designed for large-scale AI and hyperscale data centers. 

In most data centers today, power from the grid arrives as AC before being stepped down and converted to DC several times along the distribution chain. AI server clusters, particularly racks built around GPUs, consume a consistent high level of power. Its repeated AC-DC conversions reduce overall efficiency and increase thermal load. In comparison, DC-based designs, which are now gaining traction in high-density AI installations, distribute DC throughout the power system up to DC busbars of data centers, minimizing conversion stages and reducing power losses. 

Addressing this growing demand, the project targets a modular 2-5 MW SST building block that uses Infineon’s silicon carbide (SiC) switching technology combined with SolarEdge’s DC-coupled power conversion architecture. It will be engineered for direct conversion from medium-voltage grid inputs (13.8 kW to 34.5 kV) to a DC bus (800 V to 1,500 V), eliminating several intermediate steps compared to the standard alternative architecture. According to the company, the SST is expected to push power-conversion efficiency above 99%. Beyond efficiency, the design aims to optimize the electrical equipment footprint. The company adds that it will cut down on cabling and balance-of-system components, which eventually leads to faster installation. 

SolarEdge CEO Shuki Nir remarked that the surge in AI computing is forcing a rethink of how data centers are powered, stressing that reliability and efficiency are now priorities across the entire energy pathway. He said the move toward DC-centric architecture gives operators a way to cope with mounting AI datacenter energy demands. Infineon took a similar view, describing the partnership as a step toward widespread adoption of 800 V DC power systems. The company’s Chief Marketing Officer, Andreas Urschitz, highlighted that progress in semiconductor materials, Si, SiC, and gallium nitride (GaN), will play a decisive role in cutting conversion losses and in supporting long-term sustainability goals across digital infrastructure. 

Related Stories

No stories found.
logo
TaiyangNews - All About Solar Power
taiyangnews.info