When Services Start Looking Like Software

The market's obsessed with who AI will disrupt, but seems far less interested in who it will elevate — outside the obvious supply chain players.

At Energize, we're fortunate to sit in the middle of two of the most dynamic industries right now: energy and tech. On the tech side, AI is the only thing anyone's talking about. On the energy side, it's all about demand. Both are massive growth levers right now.

They're also combining to create another opportunity that's been ignored in tech circles, including by me.

For a long time I was a software purist (despite having a background in some of the best services businesses in energy). And I was a purist for good reason: software is the best business model to ever exist, especially vertical software in critical industries.

Huge gross margins. High retention customers. Low sales and marketing costs. Immense profitability.

Those traits aren't going anywhere in critical infrastructure software. If you want to know why, I'd recommend reading this post on why industrial AI is here to stay.

The short version: we'll get innovation from startups, and incumbents will get the first shot at deploying AI as the trusted brands in their space.

On the energy side of the equation, there's a business model that's now a lot more attractive: services.

Two years ago, I wrote that the world was sleeping on services businesses in energy. That hasn't changed. With AI, the conviction on that thesis is higher than ever.

Here's why.

For the first time, services businesses can have financial profiles that look more like software. Gross margins in services are driven by headcount and labor costs. Traditionally, that's capped by working on one project at any given hour.

Consulting is a good example. A consultant could only work on one deck for one client at a time, limiting their earning power for the firm. That constraint is rapidly evaporating.

With AI agents, you can work on many decks or models at once. Fire up PowerPoint with the Claude plug-in or use a standalone app like Gamma, prompt it, and let it run. While that's generating in the background, you're setting up another deliverable or taking a call to close another deal.

This ability to max out labor will lift gross margins in services businesses moving forward (something software's benefitted from since the cloud revolution).

I don't think this trend is tied to only desk-bound services, either.

OpEx across all services businesses is bound to decline. Workflows in accounting, sales support, and customer service are all getting more efficient. That goes straight to the bottom line.

And what about the revenue stickiness that software has enjoyed? That's there, too. In a lot of industries around energy and critical infrastructure, a trusted relationship matters more than anything.

Need to keep your energy bill low and make sure you're resilient to volatile prices? You call the person you've trusted for a decade to make sure that $1M bill doesn't become $2M in the summer.

Need to build an app that handles sensitive or hard to understand data, but aren't a software firm at heart? The IT services firm that understands those complexities is getting the first shot.

Need power lines inspected or designed? That contract isn't just going to anyone, it requires real expertise and a trusted name in the space.

This opportunity isn't new to private equity. Some of the best PE firms in the world have made fantastic returns in services. Subject matter expertise, pricing power, and strong profitability — even with lower growth — have usually been a recipe for success.

With AI in the mix, it's about to work even better.

So while there's a lot of concern in the market about AI right now, there's a whole category of companies out there that'll benefit from deploying it in their businesses. I'm surprised we're not hearing more about it because those companies are about to get a lot more profitable.

Ideas I'm Considering

Self-confidence is the ability to avoid the need to find signs that you are right.

Praise specifically, criticize generally.

Ideas I'm Collecting

Constraints breed resourcefulness. - Amazon Unbound, Brad Stone

Great investors are much more than fast, unemotional processors of data. They have to be strong exactly where Claude admits AI might be weakest: in dealing with novel developments where there’s not enough prior experience for dependable patterns to have been compiled. - AI Hurdles Ahead, Howard Marks