Power Play Ep 4: Solutions

Okay, so a lot of these fixers are sitting there ready to go.

What's actually stopping them? Here's what Kaie Kelly from Lower Carbon had to say.

We're talking about power and energy here, but there are transformers, there is switchgear, there are substations that go into every build.

There's, of course, interconnect cues and all of that we'll call demand on the grid.

I think a lot of the bottlenecks are actually around the equipment that goes into these different factories.

A lot of it comes down to the actual equipment needed to build a transformer or a substation.

You are familiar with the amount of lead times and backlights for turbines.

There's a very significant backlog in those places.

And so it's different parts of the value chain that are seeing some slowdown besides just end consumer demand areas.

A utility is in many ways a monolith, but they exist to provide secure power to individual consumers.

And so by nature, they're not going to be very fast adopters to technology.

What they look for is a de-risk solution and they should be relatively slow buyers.

One of the big insights we have, and you'll see this in the founders that we back, is companies who genuinely can speak the language of those grid operators and utility is a very hard contract to win.

But once you win it, you are very locked in terms of your ability to continue to work with them over time.

And so one of the enduring advantages that we have, even as an investment firm is just understanding the nuances of a space that has enormous opportunity.

but also requires a very particular type of company to build trust with those utilities in order to sell to them effectively and to build partnerships and coalitions that are required to add genuine, flexible solutions to what exists today.

Optimistic takes tend to skate past something on the demand side.

There's an old idea in economics called Jeven's paradox.

It states that when something gets cheaper and more efficient, we tend to use a lot more of it.

So every gain in how efficiently a chip runs can get spent on running even more AI, and the total power draw keeps climbing.

The efficiency gains are real, and demand for AI tends to grow right past them.

Is there this concern that if there was an AI bubble or pop, what does that potentially do to a lot of these larger, longer-term investments if that happens?

If for some reason cloud demand drops off, we realize we actually don't need it at all.

The really interesting thing there is that demand for basic internet and usage of data and compute is not just for the

purposes of inference or large language models.