TE Connectivity

Physical connectors and sensors cannot be downloaded or prompted into existence — every autonomous vehicle, data center, and manufacturing robot needs thousands of TE's tiny mechanical components to function. The business model is selling atoms, not bits, which makes it refreshingly AI-resistant in a world going digital.

Factory automation and computer vision will eventually handle connector assembly and quality testing, but TE's engineering workforce designs components for specific physical constraints that require understanding materials science, not just pattern matching. The assembly line goes first; the PhD electrical engineers designing 5G antenna arrays go last.

Forty years of manufacturing process patents, relationships with every major OEM from automotive to aerospace, and the kind of supply chain depth that takes decades to replicate — plus regulatory certifications for safety-critical applications where switching suppliers requires months of re-qualification. This is what an actual moat looks like in the physical world.

Their 'Industrial Internet of Things' strategy and edge computing sensor development suggests they understand AI creates more demand for their hardware, not less — every smart factory needs more sensors and data connectivity, not fewer physical components.

AI acceleration increases demand for TE's products exponentially: every data center needs power distribution, every autonomous system needs sensors, every robotic assembly line needs more connectivity than the factories it replaces. They're not getting disrupted — they're getting a tailwind disguised as a threat.