AI supply-chain thesis — mapping bottlenecks, focus companies, and supply-chain exposure for investors.
**Bottleneck theme:** Power & Grid **Focus:** $HUBB — HUBBELL INC Utility transformers — particularly the medium-voltage and high-voltage units that step power down from transmission to distribution and into the data-center yard — are the actual binding constraint on the AI build-out. Lead times have stretched from 6-12 months to 2-4 years, prices have doubled, and U.S. domestic manufacturing capacity is a small fraction of demand. Hubbell's Utility Solutions segment sits squarely in this bottleneck: pad-mount transformers, switchgear, grid hardware, and the integrated medium-voltage distribution products that every hyperscaler campus and every grid expansion now requires. Beyond data centers, the structural drivers are layered: aging-grid replacement (much of the U.S. fleet is 40+ years old), reshoring-driven industrial electrification, and EV charging infrastructure. Hubbell has pricing power that legacy utility-vendors haven't seen in decades — the order book is multi-year, customers are paying upfront, and the rest of the supply chain ($HPS.TO transformers, $ETN/$GEV switchgear) reinforces rather than competes for the same revenue dollar. Risks: a meaningful drop in capex from the regulated utility customer base, copper/grain-oriented electrical steel cost surprises, and competitive incursion from $ABB, $SIEGY, and Asian players if domestic premium pricing pulls in imports.
Browse all AI supply-chain theses · Macro trends · Industries · Product categories