In 2024, the total electricity consumed by data centers worldwide exceeded the total electricity consumption of many mid-sized countries. By 2027, that number is projected to double. The primary driver is artificial intelligence: training frontier models, running inference at scale, and powering the infrastructure that supports billions of AI interactions per day.
For energy infrastructure investors, this is the most significant demand signal in a generation.
The numbers are staggering
Training a single frontier AI model now requires hundreds of millions of dollars in compute, much of which translates directly into electricity demand. But training is only part of the equation. Inference, the process of running trained models to serve user requests, consumes far more energy in aggregate because it happens continuously at massive scale.
The major hyperscale cloud providers have committed to building hundreds of gigawatts of new data center capacity over the next five years. Each gigawatt requires not just the data center itself, but the power generation, transmission, cooling, and water infrastructure to support it. The total capital investment required is measured in trillions of dollars.
Where the investment opportunity lives
The obvious plays, semiconductor companies and cloud providers, are already priced for this growth. The underappreciated opportunities are in the enabling infrastructure: natural gas generation for reliable baseload power, renewable energy projects co-located with data centers, cooling technology companies, and the transmission and distribution infrastructure that connects it all.
In Asia and the Middle East, where much of the new data center capacity is being built, the opportunity extends to greenfield power generation, LNG import infrastructure, and water treatment systems for cooling. These are capital-intensive, long-duration projects with attractive risk-adjusted returns for investors with the right expertise and relationships.
The convergence of our experience
At THOT Capital, our decades of experience in energy infrastructure, from power plants and LNG terminals to renewable projects, positions us uniquely for this moment. The AI energy boom is not a technology investment. It is an infrastructure investment, and infrastructure is what we know.
AI does not run on algorithms alone. It runs on electricity, cooling, and physical infrastructure. The investors who understand this will capture returns that the technology-focused crowd overlooks entirely.
The companies that build, own, and operate the physical infrastructure behind AI will generate returns for decades. This is not speculation. The demand is signed, the contracts are real, and the construction is underway.