Starting from "Indiana, Michigan" through "data centers are the hero", I can't fully express how deeply I dislike and distrust this guy. Not just "data centers will make power cheaper" with a few alleged cases and "just trust me" it's because of data centers, with no interrogation of the cases where the correlation is inverse, and then - nothing at all on "what is the effect of data center demand growth on fossil fuel retirement and net zero progress". Just a full sidestep/whitewash. I applaud David Roberts for bringing that up, but I am - even as a complete amateur with a busy life, watching the enviro groups state in very explicit terms, this - data center surg - is a disaster for net zero. David Roberts, as a reader for 20 years, you have a responsibility to make that very clear, in my opinion, and soon, on your number one energy system podcast.
I agree with you on data centers. I think the energy used by data centers could be better used for passenger rail, which is one of the lowest CO2 emitters. We need more passenger rail in the US to reduce CO2 emissions.
I am delighted to hear that this kind of large scale battery application can increase energy output significantly with a much smaller investment than building new distribution infrastructure. HOWEVER, I agree with Jordan. I do not understand why we would benefit from using so much of that capacity for AI (which is also bad for other reasons) instead of using it to accelerate the transition to a clean energy society. If there is a good reason, I would really like to know, but I did not find it clearly expressed in this conversation. Also, will that increasing demand come from fossil or renewable sources? Also unclear. Thanks, David. Love your work!
During the last legislative session in Olympia, the Sierra Club did not say: "the data center surge is a disaster for net zero" They said to regulate data centers.
Otherwise, yes, let utilities build and own stuff on customer property, however they compensate said customers. This avoids the decades of financial engineering attempts that seem to always end up with 1% market penetration of efficiency, solar, HPs, etc., in rental property with the "split, or inverse, incentive," or C/I buildings where managers have 1000 other priorities. I mean, most businesses don't even want to own their own copiers, or these days, their own servers.
To be more specific about my question from yesterday. In the April 22 Podcast Ember told you that the price of battery packs was down to $70/kWh. Why is the Xcel utility program budgeted for over $500/kWh? How does a battery at this specific real-world cost bring rates down by 2030 when it is up and running? It would be great to have a podcast to discuss these discrepancies. Similarly, another recent podcast with Saul Griffith talked about the functionality of Tesla home batteries, which I think cost about $1000/kWh installed. Towards the end of the pod he threw in a battery cost of $90/kWh as though you can get a home battery for $90/kWh. Where does reality land on prices for actual systems in the USA? What is the price per kWh dispatched through the battery? What is the use case at each size and price point?
If you live somewhere with quite cheap power, when do batteries start to be useful beyond just replacing a few of the most expensive peaker plants?
David, I real like your podcast and have been a paid subscriber for years. I had a couple of serious questions about this one.
The podcast has a theme that batteries are cheap, perhaps nearly free! Over and over I heard that batteries are super cheap and will lower rates. Yet, the actual program that you were discussing, Xcel's 200MW of batteries at $430M, is very expensive.
Given the perverse incentives, it is also heavy in capital investment up front which is very utility-friendly when it comes to profitability.
By my approximate calculations, if you have solar at $0.05/kWh and store it in these batteries once per day to use later, you will spend the $0.05/kWh for the solar, and then $0.24/kWh to store it in the batteries. How is this cheap? The budget for this procurement is VERY expensive compared to capacity cost in other programs elsewhere. I assumed that this was 4-hour storage in this calculation, but this would be another key fact to know.
Please ask your guests about costs! Also do your own calculations in preparation for a guest likely to be advocating for their product.
It is NOT at all clear that this guest was a clean-energy advocate. Much of the grid cost savings that he was speaking of is likely using the batteries to enable the gas and coal plants to run at a higher capacity factor. Does Xcel Minnesota currently have excess renewable power to charge these batteries reliably?
In the actual filing for this program, the PUC noted that: "The Commission also agrees that Xcel was not as transparent about its relationship with Sparkfund as the Commission would have expected Xcel to be, since Xcel only revealed its relationship to Sparkfund in its reply comments".
Xcel has an investment in Sparkfund via a limited partnership. It is far from clear that you were getting a disinterested technical assessment from this guest.
Utility-scale batteries and industrial solar on the distribution grid may well be part of the solution for a renewable grid. Years ago, this was the conclusion of Chris Clack's study that "Local Solar for All Cost Less". But in that case, the assumptions for the cost of the batteries were significantly less than Xcel/Pier LaFarge are proposing here. Costs matter if we want to electrify everything.
Batteries are great but does this specific program really show the best that can be done? Does it justify the giddiness about ubiquitous batteries as THE key to low-cost clean energy?
At which point did we wave the magic wand so that utilities no longer earn excessive return on equity for capital investments? That's a precursor to cheap electricity.
As a Georgia power customer I think this guy got alot wrong about data centers and southern company in Georgia. Our psc allowed Georgia power to merge with Atlanta gas and light in 2016 so southern company owns much of the gas infrastructure. Of course they want more gas and less solar! They are delaying retirement of coal plants for data centers and incentivized to keep us hooked on their gas. We have only 7% solar. . In Fayette county Georgia power has eminent domain and is tearing down trees in their front yard for transmission lines and there is not a thing people can do about it. This is the public good? Data centers have really cheap power rates vs residential rates and Georgia power has a higher than normal ROE. To me the past public service commissioners have failed us. They accept money from the law firm that represents Georgia power during the irp. I’ve gone multiple times asking for more clean energy as others have. How is it that a centralized model is benefitting us? Our commissioners took away energy efficiency! I got solar and a battery now and my bills are cheap. That was what helped. Georgia power just lobbied against balcony solar in our state. No piece of the pie for every day Georgians! The regulated monopoly system is hampering us, not helping. I also pay an environmental compliance fee-10% of our bill for coal ash clean up. Let’s stop burning expensive coal! There is a villain. It’s southern company and the regulators who haven’t regulated. I’m looking forward to voting in one more new person this year.
Electrification of transportation is one of the biggest slices of the "crap ton" of distributed batteries Dave really wants. Additionally, many EV charging stations include buffering batteries and sometimes solar panels. Additionally, many EV fleets and individual owners install batteries and solar panels to buffer their transportation costs. My point is we should not view all transportation electrification to be new load on the grid, Some of that load will be offset by onsite electricity production and the EV's and charging stations themselves are part of the distributed grid battery infrastructure being advocated for in this podcast.
Even if Lafarge's position that distributed batteries will allow fuller utilization of the grid, which I agree with, and that allows data centers to pay to utilize the freed up capacity, which I also agree with, we still have the problem that the data centers are building explosively now, and batteries are lagging to far in installation to achieve that goal. Oh, and the fact that millions of people may lose their jobs, fascists governments will increase their surveillance of their disgruntled unemployed populace, and a host of other social ills could result from this technology, etc., all lead me to believe AI can can be postponed for a few years IMHO while batteries, and institutional regulations catch up.
“ and batteries are lagging to far in installation to achieve that goal”
On that very specific point you are making: I think the point that LaFarge and Dave both are making is that batteries can be installed at lightning speed and low cost. Or maybe and I misunderstood them because everything I’ve read lately about batteries point to that conclusion.
First the reduction in price and fast deployment of solar PV, then the same dynamic with batteries: it is just crazy. What this combo has done and poised to do is one of the most amazing things happening in the world.
Finally finished the pod. Good stuff in there; I liked the description of the "horseshoe" of utility performance or whatever, seemingly with the ERCOT-style restructuring at one open upper end, PJM-capacity market jive at the bottom, and some vertical utilities well overseen in a few states up at the other end. If I got it.
I do buy that storage of a thousand flavors is an insanely great improvement for a grid with variable loads and variable generation. (Instead of "time machine" I like "Silos for solar." I'm not convinced that it will reduce rates with all new loads in all places. Or will clean the grid in the near term. Cheap gas or coal in some places can be the "marginal" generation that feeds the batteries.
Checked the bluesky comments. Yikes.
Kinda felt like you about the "Dems attack electric utilities" Politico post. Whatever their flaws, the electric utilities have added clean energy and incented efficiency more than just about any entities over the last few decades. OK, the Fed stds for appliances, and just the LED revolution are right up there too.
I think the social contract analogy about the power system that you get as much power as you want anytime you want it is exaggerated. If you are in a rural Coop service area you may be able to get service anywhere you want it, but in most places only the first 500' is provided free. After that it will cost you at least a several dollars a foot. That is why we built off grid.
That's where I tuned out! 😄 I'll finish the episode, but that was nauseating. Also, just look up the demographics on who suffers when there are grid problems. Egalitarian? Perhaps relative in the grand scheme of all things American, but it's a point for which I have a lot of counterarguments.
We can thank this situation for some great innovations, though.
Yeah. I helped build a pal with his off grid PV-battery system under those circumstances 30 years ago. Also, even within utility territory much closer than 500' there can be onerous charges for service upgrades. Which might in some cases be unneeded with onsite batteries.
I'm inclined to agree with the argument that we should not be obsessing about AI data centers. Just require them to supply the required generation & transmission for their demand, and if they do that by co-locating with generation and batteries, so much the better. But that's all "Industrial Tier" infrastructure, with marginal immediate benefit for typical residential and small business rate payers.
If you're talking about rapidly deployed, widely distributed battery benefits for residential and small business rate payers, that's BEV batteries and Vehicle-2-Grid chargers in widely distributed solar canopy parking lots, with on-site non-flammable BESS, in large parking lots at existing workplaces, shopping centers, apartments and condominiums, everywhere. That's "Community Tier" infrastructure. In California, there are 5 silos that all need to be inter-connected with SaS to make that happen: Parking lot owners, local government, public utilities, VPP developers, and Community Choice Aggregators. And the parking lot owners need to get entirely new income from 20-year airspace leases on their lots.
China has established the world’s largest electric vehicle (EV) charging network, totaling over 21 million charging points. This reflects a nearly 48% year-over-year growth, widening a massive gap with the United States (~200,000 public stations) and the European Union (~910,000 public chargers).
Commuters require extraordinary fuel cost reduction and convenience incentives to rapidly switch to EVs. That’s exactly why France & South Korea have already mandated solar canopy parking lot Virtual Power Plants at ALL existing parking lots over 80 spaces, nationwide, within 3 to 5 years, largest lots first. That's how you electrify, reduce & stabilize transportation costs, everywhere, really FAST! Note that this strategy also rapidly reduces peak transmission grid demand for shopping center refrigeration and small business and residential air conditioning loads
Has anyone been able to find the actual citation to the very important (if true!) statement by the MPUC that LaFarge points to when he says:
"At the end of the day, the reality of a grid is that it is public infrastructure in Minnesota. This was very specific to the decision that the commission came down with, David. In this 5-0 vote, they put this intentionally and with great legal specificity. I encourage those among your listeners who are truly the most wonderful nerds to go and look at the transcript, look at the hearing. This is public record. It's a civic process. The commission said very particularly, legally speaking, the grid is owned by the ratepayers. Grid operators just take private money, Wall Street's money, they build the grid and then are given a guaranteed or partially guaranteed return so that everyone can use it. They use it by paying for power. The legal status of the grid is it is owned by the ratepayers."
OK I'm ready to say that this just didn't happen, or at least not at the April 2, 2026 meeting. There is discussion about the system's ownership at ~4:32:00 but there is absolutely NO conclusion reached and LaFarge's assertion that this was somehow definitive is just not backed up by the actual record. I'm really confused about why he made such a big deal about this and it's concerning, given his role/interest in this matter.
Starting from "Indiana, Michigan" through "data centers are the hero", I can't fully express how deeply I dislike and distrust this guy. Not just "data centers will make power cheaper" with a few alleged cases and "just trust me" it's because of data centers, with no interrogation of the cases where the correlation is inverse, and then - nothing at all on "what is the effect of data center demand growth on fossil fuel retirement and net zero progress". Just a full sidestep/whitewash. I applaud David Roberts for bringing that up, but I am - even as a complete amateur with a busy life, watching the enviro groups state in very explicit terms, this - data center surg - is a disaster for net zero. David Roberts, as a reader for 20 years, you have a responsibility to make that very clear, in my opinion, and soon, on your number one energy system podcast.
I agree with you on data centers. I think the energy used by data centers could be better used for passenger rail, which is one of the lowest CO2 emitters. We need more passenger rail in the US to reduce CO2 emissions.
I am delighted to hear that this kind of large scale battery application can increase energy output significantly with a much smaller investment than building new distribution infrastructure. HOWEVER, I agree with Jordan. I do not understand why we would benefit from using so much of that capacity for AI (which is also bad for other reasons) instead of using it to accelerate the transition to a clean energy society. If there is a good reason, I would really like to know, but I did not find it clearly expressed in this conversation. Also, will that increasing demand come from fossil or renewable sources? Also unclear. Thanks, David. Love your work!
During the last legislative session in Olympia, the Sierra Club did not say: "the data center surge is a disaster for net zero" They said to regulate data centers.
Okay. I’m not sure if that speaks to the issue or the Sierra Club. See: https://www.theguardian.com/us-news/2025/dec/08/us-data-centers
See also, and more specifically: https://environmentamerica.org/center/resources/energy-transition-at-risk/
I mean, it was a good episode. I'm not blaming Volts.
Chris Wright and Murdoch media and fellow travelers will do everything they can to increase the costs of, and demonize, batteries. Just like they have with wind and solar and induction stoves and EVs. https://reneweconomy.com.au/from-hockey-sticks-to-ev-fires-the-circular-economy-of-climate-disinformation/
Otherwise, yes, let utilities build and own stuff on customer property, however they compensate said customers. This avoids the decades of financial engineering attempts that seem to always end up with 1% market penetration of efficiency, solar, HPs, etc., in rental property with the "split, or inverse, incentive," or C/I buildings where managers have 1000 other priorities. I mean, most businesses don't even want to own their own copiers, or these days, their own servers.
David,
To be more specific about my question from yesterday. In the April 22 Podcast Ember told you that the price of battery packs was down to $70/kWh. Why is the Xcel utility program budgeted for over $500/kWh? How does a battery at this specific real-world cost bring rates down by 2030 when it is up and running? It would be great to have a podcast to discuss these discrepancies. Similarly, another recent podcast with Saul Griffith talked about the functionality of Tesla home batteries, which I think cost about $1000/kWh installed. Towards the end of the pod he threw in a battery cost of $90/kWh as though you can get a home battery for $90/kWh. Where does reality land on prices for actual systems in the USA? What is the price per kWh dispatched through the battery? What is the use case at each size and price point?
If you live somewhere with quite cheap power, when do batteries start to be useful beyond just replacing a few of the most expensive peaker plants?
David, I real like your podcast and have been a paid subscriber for years. I had a couple of serious questions about this one.
The podcast has a theme that batteries are cheap, perhaps nearly free! Over and over I heard that batteries are super cheap and will lower rates. Yet, the actual program that you were discussing, Xcel's 200MW of batteries at $430M, is very expensive.
Given the perverse incentives, it is also heavy in capital investment up front which is very utility-friendly when it comes to profitability.
By my approximate calculations, if you have solar at $0.05/kWh and store it in these batteries once per day to use later, you will spend the $0.05/kWh for the solar, and then $0.24/kWh to store it in the batteries. How is this cheap? The budget for this procurement is VERY expensive compared to capacity cost in other programs elsewhere. I assumed that this was 4-hour storage in this calculation, but this would be another key fact to know.
Please ask your guests about costs! Also do your own calculations in preparation for a guest likely to be advocating for their product.
It is NOT at all clear that this guest was a clean-energy advocate. Much of the grid cost savings that he was speaking of is likely using the batteries to enable the gas and coal plants to run at a higher capacity factor. Does Xcel Minnesota currently have excess renewable power to charge these batteries reliably?
In the actual filing for this program, the PUC noted that: "The Commission also agrees that Xcel was not as transparent about its relationship with Sparkfund as the Commission would have expected Xcel to be, since Xcel only revealed its relationship to Sparkfund in its reply comments".
Xcel has an investment in Sparkfund via a limited partnership. It is far from clear that you were getting a disinterested technical assessment from this guest.
Utility-scale batteries and industrial solar on the distribution grid may well be part of the solution for a renewable grid. Years ago, this was the conclusion of Chris Clack's study that "Local Solar for All Cost Less". But in that case, the assumptions for the cost of the batteries were significantly less than Xcel/Pier LaFarge are proposing here. Costs matter if we want to electrify everything.
Batteries are great but does this specific program really show the best that can be done? Does it justify the giddiness about ubiquitous batteries as THE key to low-cost clean energy?
At which point did we wave the magic wand so that utilities no longer earn excessive return on equity for capital investments? That's a precursor to cheap electricity.
As a Georgia power customer I think this guy got alot wrong about data centers and southern company in Georgia. Our psc allowed Georgia power to merge with Atlanta gas and light in 2016 so southern company owns much of the gas infrastructure. Of course they want more gas and less solar! They are delaying retirement of coal plants for data centers and incentivized to keep us hooked on their gas. We have only 7% solar. . In Fayette county Georgia power has eminent domain and is tearing down trees in their front yard for transmission lines and there is not a thing people can do about it. This is the public good? Data centers have really cheap power rates vs residential rates and Georgia power has a higher than normal ROE. To me the past public service commissioners have failed us. They accept money from the law firm that represents Georgia power during the irp. I’ve gone multiple times asking for more clean energy as others have. How is it that a centralized model is benefitting us? Our commissioners took away energy efficiency! I got solar and a battery now and my bills are cheap. That was what helped. Georgia power just lobbied against balcony solar in our state. No piece of the pie for every day Georgians! The regulated monopoly system is hampering us, not helping. I also pay an environmental compliance fee-10% of our bill for coal ash clean up. Let’s stop burning expensive coal! There is a villain. It’s southern company and the regulators who haven’t regulated. I’m looking forward to voting in one more new person this year.
Electrification of transportation is one of the biggest slices of the "crap ton" of distributed batteries Dave really wants. Additionally, many EV charging stations include buffering batteries and sometimes solar panels. Additionally, many EV fleets and individual owners install batteries and solar panels to buffer their transportation costs. My point is we should not view all transportation electrification to be new load on the grid, Some of that load will be offset by onsite electricity production and the EV's and charging stations themselves are part of the distributed grid battery infrastructure being advocated for in this podcast.
Even if Lafarge's position that distributed batteries will allow fuller utilization of the grid, which I agree with, and that allows data centers to pay to utilize the freed up capacity, which I also agree with, we still have the problem that the data centers are building explosively now, and batteries are lagging to far in installation to achieve that goal. Oh, and the fact that millions of people may lose their jobs, fascists governments will increase their surveillance of their disgruntled unemployed populace, and a host of other social ills could result from this technology, etc., all lead me to believe AI can can be postponed for a few years IMHO while batteries, and institutional regulations catch up.
“ and batteries are lagging to far in installation to achieve that goal”
On that very specific point you are making: I think the point that LaFarge and Dave both are making is that batteries can be installed at lightning speed and low cost. Or maybe and I misunderstood them because everything I’ve read lately about batteries point to that conclusion.
First the reduction in price and fast deployment of solar PV, then the same dynamic with batteries: it is just crazy. What this combo has done and poised to do is one of the most amazing things happening in the world.
As can soar and wind, and yet, they do not seem to be overtaking the growth of data centers
Finally finished the pod. Good stuff in there; I liked the description of the "horseshoe" of utility performance or whatever, seemingly with the ERCOT-style restructuring at one open upper end, PJM-capacity market jive at the bottom, and some vertical utilities well overseen in a few states up at the other end. If I got it.
I do buy that storage of a thousand flavors is an insanely great improvement for a grid with variable loads and variable generation. (Instead of "time machine" I like "Silos for solar." I'm not convinced that it will reduce rates with all new loads in all places. Or will clean the grid in the near term. Cheap gas or coal in some places can be the "marginal" generation that feeds the batteries.
Checked the bluesky comments. Yikes.
Kinda felt like you about the "Dems attack electric utilities" Politico post. Whatever their flaws, the electric utilities have added clean energy and incented efficiency more than just about any entities over the last few decades. OK, the Fed stds for appliances, and just the LED revolution are right up there too.
I think the social contract analogy about the power system that you get as much power as you want anytime you want it is exaggerated. If you are in a rural Coop service area you may be able to get service anywhere you want it, but in most places only the first 500' is provided free. After that it will cost you at least a several dollars a foot. That is why we built off grid.
That's where I tuned out! 😄 I'll finish the episode, but that was nauseating. Also, just look up the demographics on who suffers when there are grid problems. Egalitarian? Perhaps relative in the grand scheme of all things American, but it's a point for which I have a lot of counterarguments.
We can thank this situation for some great innovations, though.
Yeah. I helped build a pal with his off grid PV-battery system under those circumstances 30 years ago. Also, even within utility territory much closer than 500' there can be onerous charges for service upgrades. Which might in some cases be unneeded with onsite batteries.
I'm inclined to agree with the argument that we should not be obsessing about AI data centers. Just require them to supply the required generation & transmission for their demand, and if they do that by co-locating with generation and batteries, so much the better. But that's all "Industrial Tier" infrastructure, with marginal immediate benefit for typical residential and small business rate payers.
If you're talking about rapidly deployed, widely distributed battery benefits for residential and small business rate payers, that's BEV batteries and Vehicle-2-Grid chargers in widely distributed solar canopy parking lots, with on-site non-flammable BESS, in large parking lots at existing workplaces, shopping centers, apartments and condominiums, everywhere. That's "Community Tier" infrastructure. In California, there are 5 silos that all need to be inter-connected with SaS to make that happen: Parking lot owners, local government, public utilities, VPP developers, and Community Choice Aggregators. And the parking lot owners need to get entirely new income from 20-year airspace leases on their lots.
China has established the world’s largest electric vehicle (EV) charging network, totaling over 21 million charging points. This reflects a nearly 48% year-over-year growth, widening a massive gap with the United States (~200,000 public stations) and the European Union (~910,000 public chargers).
Commuters require extraordinary fuel cost reduction and convenience incentives to rapidly switch to EVs. That’s exactly why France & South Korea have already mandated solar canopy parking lot Virtual Power Plants at ALL existing parking lots over 80 spaces, nationwide, within 3 to 5 years, largest lots first. That's how you electrify, reduce & stabilize transportation costs, everywhere, really FAST! Note that this strategy also rapidly reduces peak transmission grid demand for shopping center refrigeration and small business and residential air conditioning loads
I think cheap, abundant, and still fascinating!
Has anyone been able to find the actual citation to the very important (if true!) statement by the MPUC that LaFarge points to when he says:
"At the end of the day, the reality of a grid is that it is public infrastructure in Minnesota. This was very specific to the decision that the commission came down with, David. In this 5-0 vote, they put this intentionally and with great legal specificity. I encourage those among your listeners who are truly the most wonderful nerds to go and look at the transcript, look at the hearing. This is public record. It's a civic process. The commission said very particularly, legally speaking, the grid is owned by the ratepayers. Grid operators just take private money, Wall Street's money, they build the grid and then are given a guaranteed or partially guaranteed return so that everyone can use it. They use it by paying for power. The legal status of the grid is it is owned by the ratepayers."
Here's the recording of the April 2, 2026 meeting. I'll share the citation if I can find it. https://minnesotapuc.granicus.com/player/clip/2675?view_id=2&redirect=true
OK I'm ready to say that this just didn't happen, or at least not at the April 2, 2026 meeting. There is discussion about the system's ownership at ~4:32:00 but there is absolutely NO conclusion reached and LaFarge's assertion that this was somehow definitive is just not backed up by the actual record. I'm really confused about why he made such a big deal about this and it's concerning, given his role/interest in this matter.
I like the blithe disregard for the popular, intensifying hatred of data centers. Sure, yeah, let's pretend they will be excellent....
Once again Australia is leading the way.
https://www.theguardian.com/environment/ng-interactive/2026/may/31/cheaper-energy-bills-battery-revolution-climate-crisis
hoo that sounds good hehehe