Episode 115

Joseph Kao Predictive Maintenance for Power Infrastructure with Magnetic Sensing and Edge AI

With Joseph Kao, PhD is the CEO and co founder of Magnefy
April 3, 2026

What we talked about

Joseph Kao is the CEO and co founder of Magnefy, a Stanford spin out building predictive monitoring for critical power assets like transformers, inverters, and power cables. He explains how Magnefy uses high frequency magnetic sensing, similar to an ECG for electrical equipment, to listen to the “heartbeat” of the grid and detect early anomalies months in advance. Joseph shares why this matters now: millions of transformers are aging, replacement lead times are stretching into years, and operators need earlier, more reliable signals to prevent costly outages and catastrophic events.

Show notes

Joseph Kao draws a sharp distinction between how the power industry currently monitors transformers, using dissolved gas analysis, which he compares to a blood test trying to predict a heart attack, and what Magnify does, which is more like an ECG: listening directly to the electrical heartbeat of the asset in real time. Sixty-five percent of the 70 million transformers in the US grid are reaching end of life, and lead times to replace them now stretch into years.

What we covered

  • Magnify’s core technology is a Hall-Effect magnetic sensor invented at Stanford that measures the magnetic field generated by current flowing through a conductor, translating that signal back into precise electrical information without touching the asset. The sensor is non-invasive and asset-type agnostic, meaning it works on everything from pole-mounted distribution transformers to substation equipment to inverters in solar and wind farms.
  • The specific failure class where Magnify outperforms incumbents most clearly is partial discharge and arcing, insulation failures where electrons leak from the conductor and which can eventually cause fires or catastrophic explosions. These failures develop slowly over months and occur at very high frequencies. Compared to traditional HFCT sensors, Magnify’s sensor is 10 times more sensitive, captures five times more data, and delivers nine to twelve months earlier detection.
  • AI does two distinct jobs on top of the sensing data: first, it classifies fault types and separates them from environmental noise to eliminate false positives, which is the core credibility problem with existing monitoring systems; second, it performs data fusion across multiple sensor modalities, magnetic, gas, temperature, pressure, to generate a dynamic health index and rank the entire fleet of assets by urgency, giving operators a clear prioritization for maintenance intervention.
  • Joseph’s path to Magnify runs from childhood neighborhood recycling in Taiwan with his grandparents, through a PhD in polymer science at UC Berkeley focused on biodegradable materials, to R&D at Dow Chemical, Apple, and Meta, where he led materials work on the Apple Vision Pro and next-generation AR/VR products. His time at Apple also included helping build the company’s battery recycling program, which he says connected directly to his original interest in circularity and resource efficiency.
  • As an angel investor through Cal4One, Joseph invests $25,000 to $50,000 checks in very early-stage deep tech hardware founders, looking for people who combine technical expertise with genuine market understanding. He cited Taya, a company started by three Stanford classmates, which pivoted from copper cycling through multiple iterations to an AI wearable necklace and recently closed a fundraising round with Andreessen’s accelerator.
  • His advice to his younger self was to have been bolder and pursued entrepreneurship directly after his PhD, rather than seeking the stability of corporate employment. He believes the research foundation he built could have translated into disruptive technology much earlier had he taken the risk sooner.

About Joseph

Joseph Kao is the CEO and co-founder of Magnify, a Stanford spin-out focused on predictive maintenance for critical power infrastructure. He holds a PhD in materials science and previously led hardware and materials R&D teams at Apple and Meta, including contributing to the Apple Vision Pro.


Episode 115 of the PreVetted Podcast.

Don't miss it

Listen on your favorite app