New Startup Aims to Revolutionize Century-Old Power Infrastructure Technology
For nearly a century and a half, iron-core transformers have served as the backbone of electrical infrastructure, powering everything from municipal grids to modern artificial intelligence facilities. Despite their bulky design, these devices have remained largely unchanged due to their proven reliability and the engineering principle of avoiding unnecessary modifications to functioning systems.
However, the exponential growth in data center energy requirements, combined with the increasing integration of renewable energy sources and battery storage systems, is pushing this traditional technology to its operational limits. This challenge has created an opening for solid-state transformers, which appear poised to finally disrupt the established market.
Recent months have witnessed significant investment activity in the solid-state transformer sector, with emerging companies securing approximately $280 million in funding. These advanced systems offer compelling advantages, including reduced component counts for data centers, enhanced grid stability, and dramatically smaller physical footprints for power conversion infrastructure.
Hyperscale Power, a new entrant in this space, claims its approach can achieve even greater size reductions than existing solutions. According to Daniel Rothmund, the company’s co-founder and chief executive, their system will be notably more compact than current alternatives in the market.
The startup recently secured €5 million in seed funding, with World Fund and Vsquared Ventures leading the investment round. This capital will support the development of their transformer prototype, as the company works to establish its position in the rapidly evolving market.
The solid-state transformer landscape has transformed dramatically over the past two years, evolving from a virtually non-existent market to a highly competitive sector. Notable players include Amperesand, which emerged from Temasek’s early-stage investment arm; DG Matrix, which has attracted backing from industrial conglomerate ABB; and Heron Power, founded by former Tesla executive Drew Baglino and supported by Andreessen Horowitz. Collectively, these companies have raised more than $330 million, according to industry data.
While Hyperscale Power may appear to be entering the market later than its competitors, both Rothmund and co-founder Sami Pettersson bring extensive experience with the underlying technology. Rothmund’s expertise is particularly notable, having earned his doctorate from ETH Zürich through research that included designing and constructing a solid-state transformer achieving 99.1% efficiency.
Although all solid-state transformers offer size advantages over traditional iron-core models, Hyperscale Power differentiates itself through technology capable of operating at significantly higher frequencies than competing solutions. Their system processes incoming power by stepping it up to frequencies in the tens of kilohertz range, performing voltage transformation, and then reducing it back to required frequency levels.
The importance of compact power infrastructure has grown substantially as server rack power density continues to increase within data centers. Current Nvidia server configurations consume over 100 kilowatts, while the company is already developing systems requiring 1 megawatt of power—equivalent to the electricity needs of up to 1,000 residential homes.
At these massive power scales, conventional transformers and rectifiers necessary for server power conditioning become prohibitively large. As Rothmund explains, the power infrastructure can exceed twice the physical size of the server equipment itself.
The ambitious development timelines pursued by AI companies and data center operators have made solid-state transformers increasingly essential rather than optional. According to Rothmund, the absence of these advanced power systems could actually impede the rapid scaling of data center infrastructure that the industry requires.
He emphasizes that the adoption of solid-state transformers is inevitable, with the primary question being timing rather than feasibility. The technology represents a critical enabler for the continued expansion of high-density computing facilities that power modern AI applications.