So, how big a deal is the AI build-out in terms of the, the demand crunch?

Data centers and AI are the energy story right now, and this also has an interesting kind of 10-year history.

So in the early 2010s, there was a similar panic.

It wasn't about AI.

It was about data centers and the internet and digitization, and it was, look at the rate of growth of data center electricity use in the United States, and if we extrapolate this out, you know, using a, a curve, it's gonna eat the entire grid in, in 10 years or so.

And that didn't happen because, uh, data centers and chips got more efficient, we figured out how to cool things better, and that, again, efficiency saved us from 2010 to 2019, and then, then something broke, and really what broke was the ability for chips and high performance computing and AI to get better at coding, computations, image classification, all the things that we might use a, uh, large language model now for.

And it was the performance of AI and digitization got way better than people were expecting, and all of a sudden the demand started to skyrocket.

What is the supply picture look like? How quickly are we expanding the amount of energy we're producing?

Um, what are the c- core elements of this vision?

We like to think about the electricity sector as, as our, th- the century's highways, right? All the ways that we're going to be building, uh, and, and creating, uh, n- new growth, new wealth, new opportunities in this century, uh, in the United States, it's gonna have an electricity component to it.

And whether that's manufacturing or if it's services, um, or if it's, um, other opportunities for new businesses, uh, electricity is the core component.

But yet we're still treating electricity a- and the electricity infrastructure as this side un- uh, unpaid attention to infrastructure that is almost entirely privatized, that doesn't have long-run vision, um, for, uh, recapitalization and reinvestment.

And that puts a lot of the, uh, the, the work and a lot of the cost of reinvestment on, on the rate payers.

It'd be like charging, um, you know, tolls to build the original Eisenhower H- Highway System kind of before we built it, in a way that, that would always stop us from, from getting to what we want.

So what is the, the g- you know, the Grid New Deal? It's, uh, new energy supply, new transmission, new distribution, new demand management.

Um, specifically, what agencies are you talking about leveraging, and then how do you... are you drawing in this smart AI fast lanes concept? So, walk us through it in a more concrete way.

Why'd you vote against them? Doesn't make any sense.

So I enjoyed Matthew's piece and the broader discussion around this that's been bubbling up on the internet for a while.

I think it, I think it gets one thing wrong, and I think that the political theory of the IRA was, how do you c- how do you join clean energy, clean economy, clean manufacturing, and industrial strategy in a way that creates a long-term, durable coalition to get big things done in the United States? And so it wasn't, how do we put solar and batteries and manufacturing in red districts? Um, th- those things happen because largely that's where, uh, some of the solar and wind resources are, but also where, um, uh, you know, in rural areas where manufacturing, uh, uh, plants started to, to pop up, those are, uh, largely represented by Republicans.

And so there wasn't this...

I, I, I think it's important for us to, uh, correct the record around the political theory of the IRA.

It wasn't make Republicans happy so that hopefully the- they'll, they will, um, vote for it.

It was, how do you make clean energy cheap and make it in America? And the th- the theory was, if it's good for Amer- if it's good for every district, it's gonna be good for the country.

So how do you think about the conflict between those two things?