Loading summary
Transcript69 lines
- [00:03]
A
Latitude Media, covering the new frontiers of the energy transition.
- [00:08]
B
From Latitude Media, this is Open Circuit. I am Stephen Lacy and it's just me today. Jigger and Catherine are going to be back with me next week. And as we start to wind down the year, we're, you know, we're in that seasonal reflection mode. And so in that spirit, we've got a bit of a different episode this week. We, we just passed the three year mark of when ChatGPT dropped on the world, which launched this commercial arms race for AI dominance. And it's become a magnetic force, some would call it a black hole for capital, the workforce and media attention.
- [00:44]
And the narrative of the last few years, driven by figures like Sam Altman, was intoxicatingly simple. Large language models are the on ramp to general artificial intelligence. And the only bottleneck to getting there is scale. Just feed the machine more, more chips and more electricity and it will get smarter.
- [01:01]
C
Look, I think compute is going to be the currency of the future. I think it will be maybe the.
- [01:05]
B
Most precious commodity in the world, and I think we should be investing heavily.
- [01:10]
C
To make a lot more compute.
- [01:11]
B
In the race to build these LLMs, the top tech companies have collectively teed up more than a trillion dollars in capital expenditures to expand their computing power. And obviously seeing an opportunity there, Utilities are planning another trillion dollars to upgrade the grid over the next three years, much of it to serve the appetite of AI data centers. Investors are cheering both parties along, and the entire logic of this multi trillion dollar leg of the race has been based on this belief that scale is the determining force for building human level intelligence. But three years in this narrative, at least for LLMs, is showing cracks. Ilya Sutskever, the co founder and former Chief Scientist at OpenAI, was on the Dwarkesh podcast about a week ago. This is one of the most influential AI experts in the world. And he questioned the idea of a scaling law for large language models. From 2012 to 2020, it was the age of research. Now from 2020 to 2025, it was the age of scaling. Because people say, this is amazing. You gotta scale more. Keep scaling the one word, scaling. But now the scale is so big. Like, is the belief really that, oh, it's so big, but if you had 100x more, everything would be so different. Like, it would be different for sure. But like is the belief that if you just honeydex the scale, everything would be transformed. I don't think that's true.
- [02:35]
You don't have to be an AI expert to intuitively understand What Suskiver is saying here, if you use these tools, you can feel it yourself. I'm a heavy user of LLMs. I love them. I use them for research, for testing ideas, basically everything that kind of gets me to the writing or decision making process faster. It's. And even if they never get better from here, they've brought me a ton of efficiency that I never could have imagined a few years ago. And there's no doubt these tools are getting better. For example, when Gemini 3 was released last month, it felt like a pretty significant leap. But LLMs still suffer from these common problems. You know, generic writing hallucinations, the absence of real understanding. And even after a trillion dollars in training runs in GPUs, we're still a long way from general intelligence. Three years after promising a rapid march to AGI, Sam Altman has softened his tone and repositioned ChatGPT as more of a social media like engagement product than an emergent mind. We are in the middle of a major vibe shift.
- [03:37]
So why am I bringing all this up? This is an energy podcast on an AI podcast. Well, if scale alone doesn't get us to this promised intelligence, at least for LLMs, what happens to the energy and infrastructure projections built on that assumption? The energy industry has been completely sucked into this AI moment, and in some respects a lot of the decisions being made right now, you know, the planning, the procurement, the resource forecasts are running on vibes. Utilities are getting overwhelmed with speculative data center proposals that may or may not materialize. But those requests still drive up load projections. And those projections become the foundation for 20 or 30 year investments in the grid, in generation and in rates. People are making monumental resource decisions based on a vision of scale that may not unfold the way we think it will. Now, to be clear, I'm not skeptical about the need for more compute. I'm actually very bullish on it. AI isn't just LLMs. It's deep learning systems, scientific models, robotics optimization, a whole suite of tools that I think will reshape how we discover how we invent and how we integrate clean energy. But headed into the fourth year of this AI race, there is a real tension emerging. There's obviously a question about bubble dynamics, the circular deals, the creative financial instruments, and the exposure of data centers in pension funds and REITs. Those are all starting to worry analysts. If AI revenues don't catch up to AI capex, the whole ecosystem could get wobbly. And that is not a good thing for long term energy investments. We also have industry leaders publicly questioning whether the scaling laws that justify this entire build out are actually holding. And we have an electric system preparing for scenarios where these AI companies grow relentlessly, even though no one can say with confidence whether they'll be around in anything like their current form five or ten years from now. So that mismatch between the speed of the AI industry and the longevity of the energy system is one of the defining challenges of this moment. And honestly, the story is nowhere near settled. So that is where my head is on the AI energy nexus headed into next year. And so in this episode, we're featuring some conversations from our Transition AI Conference from June with the people who are actually living inside this tension, the ones trying to build projects, secure power, finance new capacity, and make sense of shifting load signals. That is all coming right up after the break.
- [06:06]
A
NextTracker is now next power as electricity demand surges with AI data centers and electrified infrastructure, solar is the only power source that can scale fast enough to meet this moment. NexPower is the revolutionary technology platform ready to deliver it. By designing connected systems that unify the structural, electrical and digital technologies of a solar power plant, nexpower is enabling smarter, more resilient energy generation that's ready to power our electric future. Powering what's next? And@nextpower.com Open Circuit is brought to you.
- [06:37]
B
By Natural Power for nearly two decades, Natural Power has provided engineering and consulting services for renewables projects across the US Natural Power supports clients in wind, solar and battery storage with a focus on independent engineering, technical due diligence, energy estimation, and developer support. With more than 245 gigawatts of project experience in North America and acceptance from major financiers, Natural Natural Power is responsive, able to meet tight timelines, and pragmatic. Natural Power works with you to understand, quantify, and mitigate risks. Learn more@naturalpower.com or click the link in the show notes.
- [07:19]
So a few months after ChatGPT blew up, we launched our Transition AI conference series. And in those early days, the questions were big and optimistic. How could these tools help operators manage an increasingly complex grid? How could AI speed up clean energy deployment? But over the last couple of years, that tone has shifted a lot. Not because the tech stalled out, but because the energy world ran into the messy part of the story. Utilities have been pretty slow to adopt AI, interconnection cues have exploded and the conversation moved from what AI might unlock to the fundamentals of the energy market itself, financing risk, project sequencing, and the growing urgency around planning. And so this week we have Two conversations that sit right at the center of that shift. One on how developers and investors are navigating this new landscape, and one on the scramble for firm power that's reshaping resource decisions in real time. And before we get to them, a little housekeeping. We're holding our next Transition AI conference in San Francisco in April 2026. Dates and location are coming very soon, so make sure you're subscribed to Latitude Media's newsletter to get the announcement. We are going to continue to focus on the mechanics of what actually gets built. The risks, the timelines, the planning tools and the capital structures. Also, if you want a more real time conversation on those topics, our CEO Scott Clivenna writes a killer newsletter called the AI Energy Nexus. If you're not already subscribed, go fix that. You can find it on LinkedIn or laatitudemedia.com just hit the subscribe button. And finally, if you appreciate the depth of these in person conversations, check out our Power Resilience forum in Houston, Texas. It's on January 21st through the 23rd. It is going to be loaded with utilities, investors, tech companies talking about the increasingly urgent question of grid resilience. More@resilience-forum.com okay, so now onto our first conversation. This one features a conversation between Scott Klevenna and Allison Clements and Peter Nolson. Alison is a former commissioner at the Federal Energy Regulatory Commission. She now works on data centers in energy. And Peter is a manager managing director at Generate Capital who was increasingly financing the clean energy projects that serve digital infrastructure. And this discussion kind of gets to the heart of what's changed right now in this market, the way these projects are financed and the new risks that come with building power and load at the same time. Here it is.
- [09:41]
D
I left FERC in June, at the end of June of last year and it was probably about two months before that where the data center driven load surge was hitting regulators smack in the face. And I will count myself as one of them. Who was behind the eight ball on this. I mean the, the, the pace increased rapidly in a very short period of time, but we were really caught off guard and deer in the headlights and it just seemed like this is where, if you want to help facilitate the energy transition, this is where you need to be, right? You need to be in this place of helping to solve for this issue. And that's what got me interested in coming over and trying to figure it out. I think it's important to note that there's just so many types of players you know, you've got the custom, the end use customers, you've got the data center developers, you've got the land developers who sell to the data center developers. You've got every flavor and stripe of investor, you know, they want to come in and people willing to give different kinds and different terms and different risks of money. So when we have these conversations, keep in mind that it's not just the utility and the hyperscaler, there's a whole ecosystem of people who sit around in different rooms like this and have these conversations.
- [10:50]
C
Yeah, totally agree. I think for generate, we first began investing in energy infrastructure in data centers in 2017. The value proposition then though was really cost savings and load management. Fast forwarding, let's say into 2022, 2023, we invested again in infrastructure at data centers, but it became speed to power and it became connecting data centers to the grid when they otherwise wouldn't be able to connect. So off grid or bridge power type applications, similar technologies as 2017, but a complete new and additional use case for customers, which was super interesting and I think that's probably the majority of our pipeline on the microgrid side today. The other thing I'd say with respect to Allison's point, completely agree with the idea of coordinating the digital side and the energy side, the complexity sort of multiplies, right? Because I think anyone who's done a big energy infrastructure project knows it's one thing to do 500 megawatt energy project, it's another thing to build the load at the same time you're building the energy. So now you have a massive data center and a massive energy project. The EPCs, the developers, the sponsors, right. The worst, the investors and then the tenants, right. The tenant of the data center who usually is calling all the shots. If those are separate entities, you're upwards of seven plus entities. And so if we think about.
- [12:25]
The OpenAI project in Abilene, so you have OpenAI, Crusoe, Blue Owl, Lancium, Softbank, Energy, Oracle and so on. Right. And those are just the names that we know about. So I think that the speed to power, which we've talked about all day and getting folks interconnected in off grid scenarios is something that we couldn't have imagined five years ago. And then in addition, the new complexity of building both at the same time has been a great challenge and opportunity.
- [12:56]
E
Yeah, I'm sure. I'm curious of how what you're describing, with so many different parties having to come together to do this, is that reflected in any way in the capital stack, is it just debt and equity and different people hold different or does the capital stack look different? Are there different types of capital in different terms and all that? How complex does that get and how much do you have to manage that? Or does everybody just bring their lawyers and figure it out?
- [13:29]
C
Yeah, it's definitely in my scope to manage that. I think what's unique here is first of all, we're very much in the early innings of this, especially with these larger scale projects that have to get built. I think the need for development capital is very essential, especially when you think about large scale projects. I think the standard sort of project finance of tax equity, project debt and equity is there. I think the thing is you just kind of have parallel capital stacks because while there are a few cases where you're seeing the same sponsor invest in both the data center and the energy, I think especially when you get to scale, it's the specialists on each side of the equation. And so there's a lot of coordination that goes on there. And that's primarily because folks at that scale typically tend to stick to their own circle of competence and expertise. To your point, a joint venture is a great way to solve that with a data center sponsor and energy sponsor coming together. Because at the end of the day one has to the load has to know Jen's going to be there and Jen has to know load's going to be there. And also it comes down to who's going to wear the risk of delay, liquidated damages if the data center doesn't show up or vice versa. So having that type of ARM and ARM partnership, super important.
- [14:46]
D
I am seeing it in some of the transactions I'm in as more sequentially. So you can make money at three points, you can make money at the sale of the powered land, you can make money at the sale of shell ready and then you can make money again when you go vertical. And the capital costs involved at the powered land might be 2 to 3 million dollars a megawatt, but the capital costs involved in going vertical might be 12 to 13 million dollars a megawatt. And so depending, there's much more of the Wild west in the early phases. That's where like Uncle Louie is calling up about his potato farm and he wants you to develop a data center. And what we do at ASG is actually diligence all those and take, you know, we took last year 500 down to 30, that might be good sites for data center development, but there's opportunities for less sophisticated capital and for part like longer term partnerships between let's say commercial real estate holding, you know, companies with real estate portfolios and actually it would be and like a blue Owl, like some sort of fund that is willing to go along across projects with a developer. In that sense, when it gets to the, you know, the ultimate, the end use customer and the going vertical and building the buildings and putting GPU racks in them, that there's that kind of come, that goes to your, to your world out of the early stage stuff.
- [16:10]
E
So maybe let's rank risk a little bit and how you look at this because now you've added when you bring these together, the nice thing is it's a bigger project and this is a huge infrastructure project. So the potential.
- [16:25]
Is fantastic. But it does feel like those two worlds are separate for a reason. There's different risk profiles for digital and energy infrastructure. So when you go into something in the hundreds of megawatts.
- [16:39]
What are the risks you were most worried about? How do you triage those? Because I guess I think there's so many places to find risk in this right now AI itself is still in the early stages of proving that it's a long term economic driver. So there's just the overall AI market risk. But then there's all these new exogenous risks like tariffs and supply chain bottlenecks and grid interconnection and local opposition and it feels like any one of them could sort of knock a project out. So I'm curious, you could take a me John, but how do you think about risk right now and which ones.
- [17:18]
C
Worry you the most in general generate is an infrastructure investor. We typically in the three phases that Alison laid out, we are in phase three. It's most natural for us to come in at notice to proceed of construction and after. Which means that we're used to having all of the permits and the interconnection and the procurement plan pretty well teed up. So when you're building then a data center and an energy project at the same time, the biggest new risk that comes on is ensuring that the data center load will arrive. Number one, binary it will arrive and number two, it'll arrive on time. And so what that means is for us to sort of grant notice to proceed. We're looking for an executed customer tenant agreement between the data center developer and the hyperscaler or the retail customers. Right. So that's a new piece that we typically haven't had to go through before. And I think that there's a real in some parts of the industry a push To I think everyone will tell you they're not building on spec, but to avoid building on spec. Unfortunately it is happening a little bit and I think for us and for the industry it's important that we continue to not let the hyperscalers drive too much of the discussion. So I think that's one is just getting the pieces lined up with respect to the actual tenant contract on time for when we actually grant notice to proceed of the energy project. The second is what's actually in that contract. So a lot of times these hyperscalers and of course they have very strong reliability needs and requirements themselves. Both the SLAs that are required, can we actually, can the data center developer meet those? And if they don't meet those, what does it mean for our energy revenue as an infrastructure investor? And then second, what is the term of that contract? What does it look like? And you get to second and third order effects or thoughts around, well if I need to underwrite my energy asset for 20 years, I need to get comfortable. A hyperscaler is going to pay for 20 years. But if they're signing a five year contract, okay, now I'm underwriting three renewals and maybe I can underwrite one renewal, but two or three, you have to start thinking more. And then it's like, well if it's a training load versus an inference load, right. Our view is that training is a little more risky because we're seeing more evolution in how those data centers are being developed and whatnot. And so I think going down there, when you're a long term infrastructure, fixed cash flow type investor who's used to, let's say a Walmart paying you for 20 years or that's one of the new ones. So load risk, data center tenant risk, data center contract risk, all of it's manageable. I think it was earlier said today it's a lot like real estate. That's true. I think it's a little bit new for energy infrastructure investors. And then the last is if it is an AI, load the credit, right? So having a view on OpenAI, having a view on anthropic for 20 years, we haven't done a deal on that yet with them as an off taker. But it's something that we know we'll, we'll have to be thinking about.
- [20:14]
D
I think about power all day. So to me it's the risk of even after you, you know, it used to be you'd get a will served letter and everybody was happy and that was that Will serve letters don't mean anything. You're, you know, the, the risk that even if you go through the process, you get the study done, you get an electricity service agreement signed up, that you'll still get that that power will still be available and that it will be available in the way that the utility thinks it in the timeframe it will be available.
- [20:44]
E
Well, let's talk about this speed to power thing being such a priority. It does seem to open up two opportunities and one is it creates a potential price premium so you can try out things that normally aren't competitive in traditional market. So novel forms of geothermal or micro grids or fuel cells, hydrogen. There's a lot of things that can find that seem to be getting excited about this because the premium is real and then there is this urgency around time to power. So I'm curious for you guys what you're seeing actually be successful that's novel. That's not just pulling power off the grid or solar and storage and gas. So curious what examples you're seeing and any surprises you've come across so far.
- [21:35]
D
Yeah, anytime anyone says speed to power, I think of this cartoon like Speed to power with this big cloud superhero. I think actually the speed to power, it increases the exuberance and the froth in the market, but it actually boxes out some of the novel solutions. Because what I'm experiencing at the stages one and two in this development cycle is that anything that may introduce a new risk, a first of a kind, hasn't happened. Technology that the utility hasn't worked with before, any of that people want to shy away from. And I was at a, you know, it's remarkable to me, I was at a conference last week with data center developers who are transacting they want, we need 300 megawatts in Q3 of 2028. You know, in this buy market, what can you do for me? And there's no conversation about these really great lower emission technologies ultimately could provide better resilience service than a, you know, some set of diesel backup generation or even, you know, a singular gas solution. So I think there's, there's opportunity when people like generate get involved and the hyperscalers are in it for the whole time. I know that the hyperscalers want to be doing, you know, self build. In a lot of cases they're not quite there yet. And so you, you allow this, this other market to, to thrive and to continue. And I, I'm actually discouraged and not encouraged by what that effect is. I don't think that's permanent Or I think we can get over it but we haven't yet.
- [23:13]
E
So time to power could be to play it safe too and play it safe unfortunately often means fossil fuel.
- [23:21]
C
Peter yeah, I think that was really well said and I agree with that. I think my experience has been there's a pretty wide gulf between the data center developers and our community of sustainable energy infrastructure today. I think that speed is everything. I think this customer class cares about reliability and five nines type of power Number one they're used to grid type reliability and if they can't get it they will generally seek to achieve it at the closest second form I think if you step back and you say okay what is the data center market and demand look like? We at Genera kind of put it into into two buckets One is where there's a load, existing load and you have to bring gen to the load and so this is the tier 1, 2, 3 markets places like Northern Virginia, Silicon Valley, Columbus, Ohio, Dallas, Texas where will serve letters are being delayed for a long period of time and folks are saying bring a baseload highly reliable looks like the grid asset and commission me today because I have a very valuable piece of land maybe I already have a data center on there I want to develop and build something and pick a hyperscaler telling me they want it by first quarter of 2027. In those gen to load scenarios typically energy footprint constrained the answer what we are seeing is the type of off grid and often natural gas input type resource Additionally there's the, there's the opportunity where there's a data center that is more flexible in its location that's where we can bring the actual load to gen and so that's where we're seeing things like co locating with existing renewables that are grid tied or co locating with new build what we call energy parks which solar storage at times a turbine as well often grid connected, sometimes not grid connected and so in those scenarios those are more geographically flexible loads often AI training typically are still being sized for Those high reliability 5 9s but in some cases maybe a little more flexible maybe like a bitcoin load or a crypto load or an AI training load So I think that in the former case and I don't yet have the numbers on the TAM between those two but I think in the former case when you are footprint constrained you have less flexibility on what you can do we always first say okay can we put how much solar and storage can we put on this site or our development partner will say that right? And then from there it becomes, well, maybe this is a good site for a fuel cell or maybe this is a good site for a rental reciprocating engine for three years until the grid can come.
- [26:08]
And then the last thing I'll say is we're really trying to figure out and I'm curious if folks please reach out how we can do integrated on site power and cooling and so we can bring the.
- [26:20]
PUES way down efficiencies way up as opposed to being a 40% reciprocating engine getting up to the 80 90% levels to solve those type of problems.
- [26:35]
A
Next Tracker is now Next Power Electricity demand is surging due to AI data centers and electrified infrastructure, and solar is the only power source that can scale fast enough to meet this moment. Nextpower is the revolutionary technology platform ready to deliver it. Born from the leadership of NextTracker, NextPower provides connected systems that unite the structural, electrical and digital technology of a utility scale solar plant. By designing and delivering TurnKey integrated systems, NextPower is accelerating the deployment of reliable energy generation to meet the massive need around the globe. Powering what's next@nextpower.com open circuit is brought.
- [27:13]
B
To you by Natural Power. For nearly two decades, Natural Power has delivered expert independent engineering and consulting services for renewables projects across the US and beyond. Success in project transactions requires an independent engineer who's laser focused on timelines, understands the nuances of risk, and collaborates seamlessly to develop solutions tailored to your needs. Natural Power excels at working within tight time constraints while ensuring diligence never takes a backseat. With a deep expertise in wind, solar and battery storage, Natural Power delivers top tier support and independent engineering, technical due diligence, energy estimation and developer support accepted by major financiers. Their flexible approach ensures projects are built on a strong foundation powered by expertise driven by sustainability that is natural power. Find out more@naturalpower.com or click the link in the show notes.
- [28:09]
Let's move on to our second conversation. You know, one of the central questions of the moment is whether the urgency of load growth will just favor gas or whether it will create room for a suite of emerging clean firm technologies. And you heard Alison Clements in the previous conversation say that right now speed to power is favoring fossil fuels, but at the same time we're seeing a surge of real activity in nuclear, geothermal and long duration storage. So we brought together a group of people who are right in the middle of those sectors. Our Latitude reporter Maeve Allsup sat down with Mike Kramer from Constellation, who's watching Data center demand reshape power sector planning and nuclear investments. Don Owens from Fervo Energy, which is developing next generation geothermal, and Sam Simmons from Form Energy, which is building multi day storage for exactly the kind of reliability challenges the grid is now facing. And they walk through what load growth is actually doing to resource decisions and why suddenly their phones are ringing more than ever.
- [29:05]
F
And so Mike, I want to start with you because last year you signed a 20 year agreement with Microsoft to power data centers in part by restarting the remaining reactor at Three Mile Island. And at the time, a lot of the headlines that I saw, both in B2B energy and elsewhere, were predicting that this deal was going to sort of single handedly revive nuclear in the country. And so I want to start by having you fill us in. How is the project going? What's happen the last few months and how has the broader energy landscape changed since you signed that deal?
- [29:40]
G
Yeah, it's a great question, Maeve, and obviously a lot's happened since we announced the deal back in September of last year. The project's going very well. We announced back in February that everything is on or ahead of schedule, on or under budget. And the way I break it up is a couple key areas of activity at the site that's really important to us. The first is the hiring effort. You know, restaffing a nuclear plant's obviously a big effort. We've hired over 300 people out of a total that's closer to 600 at the site, which is ahead of schedule in terms of what we had planned on in order to meet the 2028 delivery year that we had talked about initially. We're in the process of training the reactor operators and the senior reactor operators, which is a long effort, but one that is well underway. And we're excited about the opportunities to get our operators trained and ready to go operate that site. Long lead equipment has been procured, so we have some main power transformers on their way that will help us make sure that we can interconnect the facility and get it started. The interconnection request has been submitted and was part of PJM's Resource Reliability Initiative that went through last month. And then a bunch of other activity continues to go on at the site in terms of inspections of the various different equipment, making sure that everything's in a pristine order. This was one of the best operating facilities in the country before it shut down in 2019. So we intend to bring it back in the same condition or better than where it was before to Continue to operate that site at the 94% plus capacity factor that it was before it shut down back in 2019.
- [31:17]
F
A schedule not that often that I moderate panels where people are touting that about their project. So that's amazing.
- [31:22]
G
I didn't want to pick on the earlier presentation where they talked about lead time for refurbishments of nuclear plants and had a five to ten year window there. But we had announced back in September that by 2028 and I'd say we're on or ahead of that schedule. So that's a year window. Hopefully we'll beat it. And I think there's a lot of opportunities when you think about 835 megawatts of Clean24.7 generation coming on within a five year window, that's very difficult to do with any other type of technology right now. So really looking forward to that coming back.
- [31:55]
F
Yeah.
- [31:55]
G
And I was there when it shut down. It was obviously a very painful day for the employees, for the industry and for the country as a whole as it relates to our emissions profile. So bringing it back is a very exciting opportunity for Constellation.
- [32:07]
F
Yeah, absolutely. And I want to dig into the workforce question a little bit later, but Dawn Fervo has also had kind of a whirlwind 12 months. One thing that I really want to dig into that Caroline Golan alluded to. She was up here earlier for the episode of Open Circuit. Some big updates with your clean transition tariff, the sort of first of a kind financing structure that you've worked on with Google and with NV Energy. Can you give us sort of the foundational overview of the ctt? What do we all need to know to have a conversation about it and what are your learnings? I think it was recently approved by regulators. So what are your top takeaways now?
- [32:42]
A
Yeah, I would absolutely love to. Okay, so for those that don't know, the clean transition tariff is a tariff that allows customers, hyperscalers or others, direct load customers to contract directly with a gen provider and go through the utility or the municipality and not have any impact, adverse impact to ratepayers. So it's actually a vital part of accelerating the clean energy transition. If you want to use clean firm power. It's also a vital part of, I think having transmission and data center providers take back a little bit of the power of speed to deployment. And it was recently approved by the California puc. It really is just rate structure. And so what have I learned? I've learned that you can actually power the AI revolution with clean firm power. And you can do it within the market rules of today. But to do that, it requires a lot of innovation and some creativity and truthfully, a lot of trust. So what we were able to do was bring not only ourselves, but Google's regulatory team as well as the NV Energy's regulatory team together and kind of revive a terror structure that wasn't actually that new. NB Energy had had something like this, a draft of this on the books that they were working on for, I don't know, five to seven years. But they couldn't quite figure out this last piece of how do we make sure that the transmission part of this doesn't adversely impact ratepayers? And they'd always kind of gotten hung up on that. And so we were able to sort of like sit down with, in true, three third party conversations with them and kind of hash that out. It was recently approved and it has allowed us to really take the learnings that we had from our first commercial project with Google, where they helped enable and show that EGS is ready now, to the next level and have a full commercial scale of 115megawatts that will be online by 2030. Yeah, so in short, that is. It's an enabling tariff structure that I think honestly every IOU should be looking at, to repeat, if they want to make sure that they are providing the power that they need now for Clean Firm without impacting their ratepayers.
- [35:17]
F
And just quickly, before we move on, what do conversations about replicating this structure look like? What can you share?
- [35:24]
A
They need to look like the load. The. The person who needs the power needs to come together with the people who are developing the power, us, and they need to sit in a room with their rate structure folks at the iou. There's really no other way to do it. That's the way it has to happen.
- [35:44]
F
It's also been a huge year for form, and I had a really hard time narrowing down what I was going to ask you about in this first question. Not least because we actually got two Latitude Media podcast episodes about form that everybody should go listen to. But there's lots of updates there, including that at the end of last year you raised, I believe it was, a $400 million Series F. There were a couple of DOE awards that I saw in there in your press room. And of course you've been vastly expanding your manufacturing footprint. So can you start just by telling us a little bit about what your pipeline looks like right now, how big it is, who's in it, what are the kind of biggest projects on your horizons right now?
- [36:22]
H
Yeah, I mean, I think first of all, building American battery manufacturing is expensive. So that's why we have a lot of money coming in from a lot of different sources. But that manufacturing pipeline is obviously supporting a bunch of projects. I mean, I think when I think of our pipeline, there's fundamentally the last 12 months hasn't changed the nature of the pipeline that much. What Form Energy makes is a multi day storage battery, which is a way to provide firm capacity. And the need for firm capacity is obviously growing because of AI, but it's existed before, existed because of EVs, existed because of heat pumps, existed because just people need power. And so the fundamentals of the need for firm power haven't meaningfully changed. They're getting bigger and faster, but not a fundamental shift. And so a lot of our pipeline looks the same. It's utilities. It's a lot of the AI load growth. You're seeing it through utilities. And then we work with developers and ipps to support those and to go work in wholesale markets. You know, I think to be honest, one of the big things about that has changed about how we view our commercial pipeline is the CTT that Fervo did because it is an enabling, I mean not a technology, but enabling regulatory structure that benefits, I think us all.
- [37:33]
F
Next question make less sense, but kind of a fun one that I wanted to throw in. And maybe Sam uses a conversation with utilities, but I'd love for you to share if you can remember the specific conversation or meeting or time period. I imagine this is before this year or when you and your company first started to realize that the data center industry or the AI boom might be one of the biggest market factors driving your growth.
- [37:57]
G
I can start. It was just over a year and a half ago at Constellation. We started extensive efforts in working with the data companies. We had a lot of insights into the activities that were going on there from our existing commercial relationships. We're a power supplier of 3/4 of the Fortune 100. And so we talked to all those folks on a regular basis to understand their power needs and certainly saw those continue to increase. And then the additional part of it was them looking to build their own data centers. So obviously the developers had been doing a lot of that. Seeing the hyperscalers get into the build question was really one that we started to explore pretty extensively. Like I said about a year and a half ago, I think it's really important to note that this industry changes very frequently. I think you've Heard that throughout the presentations today. It is not static, it is very dynamic. The way in which the activities occur and how these power solutions are going to come about will continue to morph. So I think it's one of those where we've been trying to stay ahead of the game, understand what's going on on the customer side, understand what's going on the developer side and really think through these power solutions specifically as it relates to nuclear. And I know we talked about Three Mile Island. I'll try not to monopolize the time, but it's important when we think about what are the solutions needed for these data centers. There is a firm clean power need there and just understanding how you meet that with the different types of technologies that are available. Nuclear is a really good fit. 24. 7 has the emissions profile. You can scale it. You know, building new has some challenges and we'll talk about that in a little bit. But I think there's the opportunities to add resources to include the existing and factor in the fact that these existing plants have relicensing periods that end. And so to the extent you can extend them like we did at Clinton with the metadeal, you're really starting to see real impacts on the system for an extended period of time.
- [39:58]
A
Okay, my, my answer to this, I have two I have a personal and I have a professional. So the, the professional was actually all the way back in 2001, 21 in our early conversations with Google and Adam for me, I was having a conversation with him and we were talking about what happens next after we did our geta this geothermal acceleration development agreement that helped enable our technology. And we were talking about our future pipeline, Fervo's future pipeline and what were things we're going to like what interest Google had in the future pipeline. And I could not quite rectify what I was hearing from them with.
- [40:39]
What I was hearing and seeing from just what I could do publicly about what Google was, was interested in. So I had this like sneaking suspicion. I'm like what, what, what is going on? I am missing something. Why are you so interested in this? So that was my professional like first tip off. My personal tip off was kind of professional as well. I was actually trying to hire someone for a project manager role at Fervo. And I will not reveal the person but I, I could not get this person interested because of they were skilled at developing data centers and hyperscaling facilities. And I just could not get this person to like call me back and I'm like what is A good friend. And I'm like what is going on? Why are you not doing this? So that was in 2023.
- [41:36]
F
So Sam.
- [41:37]
H
Yeah, I mean I won't claim that we like like saw chat GPT coming before anyone else did. So I think we like we felt that chaos of like that holy crap, what is this thing and what's it going to do to the world? And then like I think you know we're a venture backed company so I think push really came to shove when our investors started calling us and be like, hey, there's a lot of money here. And like and it really changed the conversation around scaling quickly and the market potential and like how do you deploy, you know, less incrementally and less like I mean because like as a, as a technology company, you know the typical thing is to go in increments, right? Do the like we have a one and a half megawatt project coming online n later this year. That's our first one, then 10 megawatts. And then you start to get to the grid scale and it's like well people want gigawatts of this and they wanted it yesterday. And so like we started feeling calls from like investors who like were feeling calls from utilities and utilities directly saying like hey I can I get like hundreds of megawatts of this stuff next year? No, no, we just broke ground on our factory. We can't. And so I think that was the wake up call that really drove a huge acceleration at the company.
- [42:40]
F
You've outlined some of the challenges, but you also just mentioned expanding more quickly than planned. Are there certain, whether it's tech challenges or market challenges, that the AI boom, people are probably very sick of that term by the end of the day that that has smoothed over for you. Are there roadblocks that the massive demand that you're now seeing is helping clear the way for you to grow more quickly to raise more money.
- [43:07]
H
The first thing that comes to mind is just the conversation has shifted. I mean like when we, when we started trying to sell multi day storage and talking about like firm capacity and like, you know, LCOE is not a good metric because it doesn't value different hours and that sort of thing. And like what does it mean to be firm? We were like yelling into a void. And what I think this has done is everyone is now focused on what does it mean to be firm, how do I deliver firm capacity. And as you know, when you're a hundred hour battery and you're competing against four and eight hours storage, a lot of the times. That's where the difference lies is in how firm you are. And so that conversation, changing the awareness of what it means to be firm I think has been a huge change for us.
- [43:48]
A
Yeah, I would definitely echo that. I think I stopped talking about am I, do I have cost parity with solar plus a four hour battery to the speed of deployment of my pipeline. So I'd answer that question and like it did help. It helped me no longer have to stay like have all these like silly examples. I used to trying to like in panels like this talk about like what does Firmin if. How if you want to smell nice for four hours versus smelling nice all day long, you buy the more expensive perfume versus like the other. But like body Axe body spray. I would had to like make this point like over and over again. I stopped having to use that example. But the other thing that it helped me do was it helped me get in the room and have creative conversations both with off takers about early funding as well as with some of the finance community were more willing to come to the table and have conversations with me about, okay, we are a new technology, but what do you need to see to show that I am bankable? And I think that's 100% driven by what they're seeing in the projections of on spending from hyperscalers and large AI providers. I don't at all pretend that I was able to do that. They started taking our calls.
- [45:02]
G
For me, it's twofold. One is less on the demand side, but more on the urgency side. And as you think about it, we have to meet this moment. And we heard a little bit about this earlier. Jiggers comments aside, how important it is for us to win the AI race, national security, economic transformation, all the things that Stephen mentioned that we can potentially accomplish with AI. And I think everybody is on that same page that we really need to be the winner here. We can't just sit back, let other countries take over the AI conversation. And so that develops a sense of urgency in terms of how you can get things done. Now does that actually get policies in place and the results that you need? That's what we're working on. But that that sense of urgency, the need to meet the moment, I think is driving a lot of the conversation. Maeve. And then the second thing is it puts more options on the table. You know, restarting a nuclear plant is not something that people thought was feasible two years ago and now we're doing that. And we're not the only ones adding new Nuclear at certain places and making sure the existing have 20 year license lives. Not something that people were looking at that we'll buy your power for two years, not for 22 years ago. And now you have 20 year PPAs being signed that keep these units around for a long time and then the ability to add new. The new nuclear question is one we have to find an answer to. And so how do you pair these different types of products together? So you have existing taken care of, you add some new with upgrades and restarts, you think about where you can do new build on the nuclear side and you accelerate the technology development going from first of a kind to nth of a kind because of this level of urgenc. So it all kind of pairs together, it works really well. And I think it's something that probably wouldn't happen if you didn't have that sense of urgency in the amount of demand that you're talking about.
- [46:56]
H
I also just add one more thing. I do feel like, I don't know if I just certainly didn't mention it, but it is real. It created the economic white space, right? Like before we were talking to technology providers about clean firm 247 power and the willingness to pay a premium was like, like, you know, like that, like yeah, we want it. It's like yeah, but you want at the same price and now there's a willingness to pay for it, right? Like I, I don't know, you probably know this better but what did a nuclear plant just contracted like 80 bucks a megawatt hour or something.
- [47:27]
G
That's the, the, the analyst, they didn't disclose the terms on that deal that D did with, with Susquehanna. But that's the analyst estimate. I just a minor correction. When you say premium, it's paying for the attributes, right? The attributes that these types of generation sources provide. The firm, the reliable, the clean I think is really important because that is something that has value that hasn't been accommodated by the market before. And now you're seeing the market start to value that and price that into some of the deals.
- [47:56]
A
Oh my gosh, I could not agree more with that. Honestly, like I feel like we've been trying to talk about the marginal value for so long and that was our way and our method to try and say like no, these are the valuation of the unknown unseen things, the things you're not actually valuing. And from like our perspective, just to bring it back to your question, Maeve, we, we ended up like seeing that people now are valuing that somewhere between 15 and $20 per megawatt hour. Like they finally gotten that. That is actually the value of that capacity factor of the clean firm.
- [48:31]
F
My very last question for all of you, sort of a fun quick hit question. If there was one thing we could do tomorrow that would make it easier to build new clean firm power, what would it be? And for reference on a panel a couple of weeks ago, I asked a similar question to a data center developer. How could we build data centers faster in Phoenix? And his fun answer was fly several helicopters full of cash over Phoenix and drop the cash out over the city. So if that's where you want to go with this, I give you full permission.
- [49:02]
H
Sam, I feel like the right answer is like put a copper plate everywhere so you can just. But like then we'd all get be electrocuted. So I'm still figuring out the logistics.
- [49:11]
A
Of that TSA pre check for firm clean power. That's my answer. I travel a lot.
- [49:18]
G
I actually think there's two. The first is policy certainty. And that's the answer I gave on a similar panel question last week. And that sounds very generic, but the idea is you're making investments in 60, 80, 100 year type assets and not knowing what's going to happen 10 years from now makes it very difficult. And you can get offtake agreements that go out 5, 10, 15, 20 years. But the idea is we want this to be a long term investment that serves the people and the grid and the data centers and everybody else for an extended period of time. And so if you don't have that policy certainty, it's very difficult. And the second one is just the spamming of the interconnection queue is something that just it is a prerequisite for any of these projects to get off the ground. They have to go through that process. But there's so much noise in those numbers that makes policymakers and regulators very uncomfortable. And I think we all know that that's not the right number. There's definitely demand out there, but making it at scales of 30 gigawatts instead of 5 causes all sorts of problems in terms of the planning process that we can't solve easily on our own.
- [50:31]
A
So you like my TSA for preaching?
- [50:33]
G
I do actually.
- [50:33]
H
And my copper plate idea.
- [50:35]
G
I know we're a team here.
- [50:37]
A
We're a team.
- [50:38]
F
Mike's got a message. He's sticking to it.
- [50:45]
Thank you so much. Really, really great conversation.
- [50:53]
B
That's going to do it. Kathryn and Jigger will be back with me next week. Stay tuned for more details on transition AI 2026 in San Francisco in April and get your path to the Power Resilience Forum for January in Houston, Texas. Of course, subscribe to this podcast anywhere you get your shows, and if you like what we're doing here, please give us a rating and review. Open Circuit is produced by Latitude Media. I am Stephen Lacy, your host and executive editor at Latitude. The show is edited by me. Sean Marquand is our Technical Director, Anne Bailey is our senior podcast editor, and we will catch you all next week. Thanks for being here.