The market has spent years chasing GPUs, networking and data-centre operators.
But building a huge AI data centre requires something much more basic — electrical infrastructure.
Transformers, switchgear, substations and grid connections can take years to plan and install. That creates an interesting second-order investment theme: companies that manufacture the equipment needed to actually deliver electricity to these massive facilities.
Names like GE Vernova, Eaton, ABB and Schneider Electric give investors different ways to access that infrastructure spending.
What I find interesting is that this isn’t entirely dependent on which AI model wins.
Whether Nvidia, AMD, Google, Microsoft or Amazon captures more AI demand, the data centres still need electricity.
That could make the “picks and shovels behind the power” an underappreciated part of the AI infrastructure story.
The big question:
Does the next major AI investment opportunity sit one step behind the AI companies themselves? 👇
Comments