Someone Has to Invest in the Grid. Why Not Data Centers?

This week’s conversation is with Jane Flegal, esteemed energy wonk extraordinaire and friend of Heatmap News. I reached out to Jane because she recently authored a paper for a think tank – the Searchlight Institute – focused on how to try and get transmission built to satisfy growing electricity demand without creating the cost pain points that foment discontent on the ground. Y’know, how to avoid the sorts of frustrations we chronicle here at The Fight! So ahead of reporting on transmission conflicts I have coming up next week, it made sense to have a candid conversation about just how hard all of this is.

The following transcript was lightly edited for clarity.

How much of this transmission build-out needed is because of data centers?

We have underinvested in the kind of transmission and grid infrastructure that we need to grow the grid and power basically anything new. We’re seeing regulators and reliability analysts flagging some major concerns. Beyond investing in new capacity, we’re just at the 50-60 year point in an infrastructure and investment cycle. A lot of what we have was built in the 1960s and 1970s. Even if we didn’t grow the grid, there would be significant investment required in our existing infrastructure just to maintain and fix it.

I actually think even if data centers were not on the horizon at all, there would be real concerns about who and how to pay for reinvestment into the grid. The question of what this growth requires for the grid, most of the analysis mapping out what we need to do to decarbonize is that we’ll need to 2x or 3x the grid to electrify everything.

When you drill down into it, the utilities were going to need to build some of this stuff anyway. There was going to have to be huge transmission and distribution investments, regardless of data center load growth. Wildfire hardening in the West. There’s deferred maintenance coming due.

It’s also true we did not anticipate the quality of demand data centers represent and it’s so sudden and so big. The demand is so centralized. It’s a different shape of demand for what we expected for electric vehicle infrastructure, for example. It’s unique.

Then there’s the question of what’s attributable to this kind of large load growth. What’s the incremental investment that wouldn’t have been made but for these data centers? If it’s a big new transmission corridor to reach a data center campus, we don’t necessarily want those things to be socialized across the rate base. So you see multi-billion dollar transmission plans in some states where the utility or a state government will say this is due to data center demand, so it’s hard to separate those things entirely.

But what I find frustrating about the affordability conversations is these are investments we would need to make anyway and/or would be societally useful even if the data center doesn’t materialize. Not to mention that we haven’t totally figured out how to deal with that! If the assumption is that no new infrastructure is good or desirable, that’s not good. That’s bad.

The question is, who pays? Funding things through the rate base is super regressive. Electric bills represent a higher share of low-income earners’ income and so it's not a good way to fund big things. A meta question is, who should be paying for all this stuff? The data centers should pay for what they created and are demanding.

It feels like what you’re getting at here is the need for some financing backstop to blunt the impact on ratepayers. The local folks, people who don’t see how transmission will make their lives easier.

I think what I’m trying to resolve is, you need to have a mechanism to make needed investment in transmission infrastructure investable without socializing all of the cost.

Right now we’re in a lucky position because we have large customers with capital and a willingness to spend it for speed-to-power. They can help on this front both by engaging in take-or-pay commitments where they commit legally to being the offtaker and by doing up-front financing themselves in the transmission. This is a real challenge though, which is why I was trying to think creatively.

As you said, transmission investment if planned well and permitted on time can make things cheaper and more stable over time. But the investment has to happen and be paid for somehow. This has always been an issue.

I was speaking with an environmentalist in Virginia earlier this week about transmission. This is someone who doesn’t want to build a lot more transmission explicitly for data centers. So I raised the question of, weren’t we just talking about how we need more transmission for the climate? Why are you against these projects then? And what this person said was that the transmission for data centers was eating up utility funding that could go to renewable energy and could power other demand sources.

Is the question that utilities are spending on this stuff to satisfy data center demand and therefore won’t be investing in projects to power our lives? Or is it more complicated?

It’s a fair concern here and it goes back to our planning processes. If you build a transmission corridor for a data center in Virginia, that's different from a high-voltage line from the wind farms in the West to load centers in Chicago. I see what they’re saying. But the truth is the U.S. needs dramatically more transmission for electrification no matter what. The grid cannot accommodate the decarbonization required and we can’t move power from the best resource centers to load centers. That was all needed before the hyperscalers started building.

The data center build-out is an accelerant bringing forward all this investment that is already needed. If it is planned correctly it can help electrification goals simultaneously. And the “if planned correctly” part does a lot of work.

But are tech companies investing in the transmission?

They certainly are. But it's another area where we haven’t made it particularly easy for them to do that. They’ve committed to spending quite a lot of money on infrastructure but most of it is not grid. Google is investing for example into advanced conductors onto the grid, which is a shared investment that’ll benefit the public. To date however, most of the hyperscale investment is the requirements for their own load, not system contribution. That’s what I was trying to propose in my paper.

Voluntary pledges are not going to be enough. But can you get a state to condition tax benefits for data centers on a set of conditions, like dedicated capacity payments. Ideally some mechanism to invest in the broader grid. It’s a big ask of them though, it's worth saying.

Right now the barrier is we can’t plan and permit the lines to begin with, so there’s nothing for them to invest in, and my biggest concern is them just going behind-the-meter.

I think the thing that’s important here is that there’s a set of questions around what data centers can do directly with their capital and a set of questions around the policy and regulatory agenda for the grid. What I’d say is we’re having an active debate on the Hill right now about federal permitting and as a part of that conversation, we're talking about transmission. We’ve tried to do a better job at this and repeatedly failed, partially due to opposition from utilities and states at a time of flat or declining demand.

That is changing; we have large, powerful customers with a lot of money and political power who can advocate for the permitting reform we need to solve structural issues here. I think now is the moment where we have the political coalition to do this. We were never going to solve this by having climate advocates yell at FERC.