Nobody builds a Super Bowl ad around a liquid cooling system. But somewhere behind every AI chatbot answer, every image generation and every model training run, there's a data center generating enormous amounts of heat.
Without the right infrastructure, all that computing power has a problem.
That's where Vertiv comes in.
The Ohio-based company makes power management and thermal cooling systems for data centers, the kind of infrastructure that rarely gets mentioned alongside Nvidia, OpenAI or Microsoft but is becoming increasingly important as AI workloads push server density higher.
As AI computing has exploded, so has demand for the equipment that keeps those servers powered, cooled and running reliably.
The Unsexy Side of AI Infrastructure
Training and running large AI models requires enormous computing power, and enormous computing power generates enormous heat.
Traditional air cooling becomes less practical as server racks become denser and more powerful. That has pushed data center operators toward liquid cooling and increasingly sophisticated power-distribution systems, exactly the kind of infrastructure Vertiv supplies.
The company occupies a position that resembles the classic "picks and shovels" businesses of a gold rush.
It doesn't need to determine which AI model wins.
It doesn't need to know whether Nvidia, AMD or custom accelerators ultimately dominate.
It needs data centers to keep getting built and operators to keep upgrading the infrastructure inside them.
And right now, the AI buildout is creating enormous demand for both.
Hyperscalers including Microsoft, Amazon, Google and Meta have continued committing vast amounts of capital to data center capacity as they compete to expand their AI infrastructure.
That spending creates opportunities beyond the companies producing the processors and models.
A Business Investors Are Starting to Notice
Vertiv's stock performance has attracted increasing attention as investors look for companies positioned to benefit from AI infrastructure spending without having to predict which application or model ultimately wins.
That's a different proposition from betting directly on an AI software company.
It's an infrastructure bet.
Vertiv's exposure to the data center market isn't new. The company has supplied power and cooling equipment to data centers for years, long before generative AI became a boardroom priority.
What's changed is the scale and urgency of demand.
Modern AI systems require increasingly dense computing environments, increasing the importance of thermal management, power distribution and other supporting infrastructure.
That creates a potentially attractive position for companies supplying the physical systems required by almost every major AI deployment.
It also comes with risks. Data center construction can slow, hyperscalers can reduce capital spending, competitors can gain market share and investors can price future AI growth too aggressively into infrastructure stocks.
The "picks and shovels" thesis isn't immune to the AI cycle.
Why the "Boring" Layer Matters
The AI boom has produced obvious winners.
Nvidia supplies the accelerators. OpenAI and Anthropic develop frontier models. Microsoft and Amazon provide enormous amounts of cloud infrastructure.
But all of those layers ultimately depend on something much less glamorous.
Every AI data center needs power delivery, cooling, thermal management and backup systems.
Those requirements don't disappear simply because one model becomes more popular than another.
That is what makes companies like Vertiv interesting.
Their role is easy to overlook because the infrastructure isn't visible to consumers. Nobody opens ChatGPT and thinks about the cooling system keeping the servers online.
But from an economic perspective, that invisible layer can be just as important as the technology people actually see.
The broader lesson is worth remembering as the AI investment story matures: the companies making money from AI won't necessarily be the ones building the most recognizable products.
Some will be supplying the infrastructure that everyone else needs.
What This Means for Miami
South Florida isn't a hyperscale data center market on the scale of Northern Virginia or Texas, but Miami's growing technology, cloud and data center ecosystem still has a stake in the infrastructure buildout.
Local companies building or leasing data center capacity need the same power and cooling systems as facilities anywhere else.
For Miami's investment community, increasingly active in technology and venture capital, Vertiv offers a useful example of the broader AI "picks and shovels" thesis.
Investors don't necessarily have to identify the next winning AI application to gain exposure to the sector.
They can also look down the stack.
The companies supplying power, cooling, networking, memory and physical infrastructure may benefit from AI growth regardless of which model or application ultimately dominates.
That's the less glamorous side of the AI boom.
It may also be one of the most durable.