AI is moving fast. Can power and infrastructure keep up?

My interest in energy started with a simple question: how are we going to power all of this?

As I followed the growth of AI, I became increasingly interested in the physical systems behind it. I already cared about climate and sustainability and was sharing weekly spotlights on companies in the space. University courses on green technology and venture investing, alongside working in venture and conversations with founders, had taught me to look closely at the business model behind a promising technology.

But AI brought a new urgency to the energy question. I started looking into the companies working to meet that demand, and as I learned more about the power needed to support AI, I realized energy was part of a much larger picture.

Even if we had enough reliable power, would we have the infrastructure needed to support AI at this scale?

Securing the power is only part of the equation. We would still need to design and manufacture more efficient chips, build data centers, install cooling systems, and develop the networks connecting everything together. The reverse is just as true. We can build advanced chips and enormous data centers, but without reliable power to operate them, it does not work.

That is what makes this space so interesting to me! AI's growth requires progress across several systems at once. A breakthrough in one does not remove the need to build the others.

Often, the opposite happens: solving one constraint exposes another. Packing more computing into a facility can create new cooling and power requirements. Building more data centers increases demand on energy systems.

This is where I think some of the most interesting investment opportunities of the coming decade could emerge. For an investor, understanding how a company fits into the wider system becomes essential. Its technology might work, but what else needs to be in place before a customer can use it? Does it remove a constraint or depend on another constraint being solved first?

Of course, timing is another important factor. With several parts of the system evolving at once, a company's solution must align with where the market is heading. Something that makes sense now may not in five years, while something we will need in five years may struggle today.

I want to understand which companies can turn today's constraints into lasting businesses. What makes their solution valuable? What would persuade a customer to adopt it? And would that customer still need it if the immediate shortage eased?

My interest in sustainability brought me toward energy. AI made me look more closely at the infrastructure supporting growth. Now, the investment question is where that system needs to change, and which companies can build an enduring advantage by helping it do so.