TL;DR

PowerMatrix delivers highly efficient and compact power supplies for cutting-edge applications—from GPUs and data centers to drones, EVs, and aerospace. Based on proprietary technologies, our power supply systems can reduce energy loss by up to 50%. We’ve built working prototypes and are piloting with world-leading partners.

The Problem

AI computing heavily relies on data centers where every watt of electricity is processed and converted by multiple stages of power supplies—from the power grid to processors.

Currently power supply solutions face significant challenges:

• Inefficiency: Data centers consume 5% - 10% of global electricity, while 20% of that (100TWh) is wasted on power supplies.
• Size Constraints: Power supply is given very limited space in server trays and on GPU board surface but required to provide much higher power.
• Performance Gaps: It is required to provide a voltage conversion ratio over 100 and provide 1A/ns dynamics for robust operation of future advanced AI computing processors.

Our Solution

Based on proprietary technologies, we provide the reverlutionary single-stage vertical power delivery solution:

• Single Stage: directly convert high voltage to <1V low voltage, eliminating inefficiencies multiple stage power conversion chains.

  

• Vertical Power Delivery: minimize power transfer loop, locating directly underneath computing chips.

  

Results vs. incumbents:

• 80% smaller form factor
• 50% less energy wasted
• 10× faster dynamic response

Deploying our modules can save customers $10 B annually, cut 30 TWh of electricity waste, and avoid 100 Mt of CO₂ emissions each year.

Team

We have more than ten years of experience in power electronics and are dedicated to solve the most challenging tasks in this area.

Borong did a Postdoc at the University of Cambridge and got his PhD in power electronics from the University of Warwick. He has extensive experience in research: he published 60 papers that got 650 citations and received the IEEE ECCE William Portnoy Award. He also worked on a power converter for the largest wind turbine at GE, and led a team at Bitmain, a bitcoin miner, to develop a new power supply.

Xufu got his PhD in power electronics from the University of Cambridge. He received the Semikron Danfoss Young Engineer Award in 2025, the Trinity Bradfield Prize Award in 2025, and First Place Award of the IEEE ECCE Best Student Project Demonstration in 2023. He also published 15 papers on power electronics that have 41 citations.

Our Ask

If you’re building AI hardware, data-center infrastructure, or next-gen power-electronics, let’s connect! info@pwrmatrix.com