Power Electronics AI Agent

The AI that designs the power systems that power AI

We are building a Power Electronics AI Agent that closes the loop between intelligent system design and the energy infrastructure that fuels it — from semiconductor device to converter to grid.

Design across every scale

Our AI Agent operates across the full power electronics stack — three tightly connected layers that together form a complete, closed-loop design system.

Devices

Device

Characterize, model, and select semiconductor devices (SiC, GaN, Si) with AI-driven loss estimation, thermal profiling, and datasheet-to-model automation.

Learn More →
Converters

Converter

Explore topologies, size magnetics, synthesize control loops, and optimize multi-objective trade-offs — all orchestrated by the AI Agent in a single design pass.

Learn More →
Systems

System

Architect microgrids, DC distribution networks, and hybrid AC/DC power systems where every converter is co-optimized for the mission profile of the whole grid.

Learn More →

The closed-loop vision

We believe the future of energy and AI is a self-reinforcing cycle.

AI designs power systems

Our Power Electronics AI Agent assists engineers in designing optimized converters and grids — faster, with better trade-offs, and with built-in validation workflows.

Power systems fuel AI

The efficient, reliable power infrastructure that results — microgrids, DC distribution, high-density converters — powers the data centers and compute that run AI itself.

AI upgrades power systems

Every design cycle feeds data back to the AI Agent. It learns, refines its models, and produces even better designs — a true closed-loop ecosystem.

What the AI Agent does

A modular intelligence layer that plugs into your design flow — or we run it for you as a service.

Topology exploration

Constraint-aware search across LLC, DAB, multi-level DC-AC, interleaved buck/boost, and more.

Device & magnetics selection

Automated WBG device picking, core sizing, winding optimization, and switching-loss estimation.

Loss & thermal modeling

Datasheet + physics-guided surrogate models produce efficiency maps and thermal limits.

Control synthesis

Phase-shift, TPWM strategies, current/voltage loops, and soft-switching regions — tuned automatically.

Multi-objective optimization

Efficiency, power density, cost, temperature, EMI — Pareto trade-offs with clear rationales.

Validation planning

Auto-generated HIL/SIL test matrices, data capture scripts, and repeatable lab workflows.

Our Ecosystem

Industry partners building the closed-loop future of power electronics and AI together.

Panxin Technologies

Panxin Technologies

Sumitomo Heavy Industries

Sumitomo Heavy Industries

Academic Partners

Collaborating with leading research institutions to advance AI-driven power electronics.

Cardiff University

Cardiff University

Shanghai University

Shanghai University

Loughborough University

Loughborough University

University of Glasgow

University of Glasgow