Edge AI & IoT

Sector thesis

Edge AI & IoT is about running artificial intelligence directly on devices—phones, factory sensors, cars, home gadgets—rather than sending all data to distant data centers. It's the opposite of cloud computing: the intelligence lives at the edge, closer to where the action happens. Why now? Three forces collide. First, AI models are getting smaller and faster, so they can actually fit on a chip in your pocket. Second, privacy and latency matter more—you don't want to upload your medical data or wait for a cloud response when a self-driving car needs to brake. Third, the sheer volume of IoT devices (billions of sensors globally) makes it economically wasteful to send everything to the cloud. This is a structural shift, not a fad. The sector breaks into three overlapping pieces. Semiconductors are the foundation: specialized chips designed to run AI models efficiently on low power (companies making these processors). Software & frameworks are the tools developers use to build and deploy these models on edge devices. And then there's the infrastructure layer—the platforms and services that help manage, update, and monitor millions of edge devices in the field. The biggest risk is fragmentation. There's no single standard yet. Different industries (automotive, healthcare, manufacturing) are solving this problem separately, which means winners aren't obvious. Second, the economics are brutal—margins compress as competition heats up. Third, software bugs on edge devices are harder to fix than cloud bugs; a bad update could brick thousands of devices. For a retail portfolio, watch companies that sell the chips, the software tools, or the management platforms. Look for revenue growth in industrial IoT and automotive sectors—those are where edge AI creates real economic value, not just hype. Avoid betting on any single standard winning. Diversification across the supply chain is safer than picking one player.