Soluna Holdings pivots to AI data centers using stranded renewable energy with Microsoft-hired leadership
Summary
John Belizaire explains that Soluna Holdings (SLNH) co-locates green data centers at renewable energy sites to utilize stranded power, reducing interconnection timelines to 1-18 months compared to the traditional 3-5 years. The company acquired the 150 MW Briscoe Wind Farm to control power supply and accelerate AI campus development, partnering with Applied Digital for the Dorothy 3 project to share capital costs. Belizaire highlights the hiring of three Microsoft alumni with over 55 years of combined hyperscale infrastructure experience to lead the transition from Bitcoin mining to AI data centers. Investors are advised to monitor lease announcements, the closing of the joint venture, and new projects reaching Notice to Proceed as key milestones.
Episode sections
- 0:00 – 4:46
Introduction and Business Model
Tim Gerdeman introduces Soluna Holdings (SLNH) as a developer of green data centers for AI and Bitcoin mining that co-locate at renewable energy sites to utilize stranded power. John Belizaire explains that Soluna integrates directly into existing wind/solar substations to access wasted energy and grid power, reducing interconnection timelines to 20% of traditional projects (1-18 months vs 3-5 years).
- 4:46 – 10:05
Company History and Origin Story
Belizaire traces Soluna's origin to a 1 GW wind project in North Africa where curtailment was an issue, leading to the development of a Bitcoin mining solution. The strategy pivoted to the US during the pandemic, identifying similar curtailment problems in Texas with major Independent Power Producers (IPPs).
- 10:05 – 15:54
Behind-the-Meter Economics and Curtailment
Belizaire illustrates how ERCOT curtails wind farms, leaving turbines idle. Soluna's data centers absorb this curtailed power, allowing the wind farm to double its revenue by selling previously wasted energy. The model also allows Soluna to acquire wind farms outright to control the speed of converting power to compute.
- 15:54 – 22:28
Texas Political Landscape and Permitting
The current Texas administration paused data center permits and audited ERCOT's 400 GW of power requests against an 80-90 GW load. Soluna benefited from being classified in a 'firm power' bucket due to energized projects, while competitors faced delays. The new administration signals a desire to weed out speculation and ensure data centers are good community actors.
- 22:28 – 27:33
Acquisition of Briscoe Wind Farm
Soluna acquired the 150 MW Briscoe Wind Farm in April to gain sovereignty over power and accelerate the conversion to an AI campus. Ownership allows total control over the capital stack and interconnection, enabling construction of the first AI data center phase less than a year after acquisition.
- 27:33 – 33:57
Applied Digital Joint Venture and Financing
Soluna partnered with Applied Digital for the Dorothy 3 project (50 MW) to share capital costs and leverage the partner's experience for debt financing. Belizaire notes AI data centers cost tens of millions per MW, requiring a capital stack of 70-80% project-level debt and 20-30% equity, targeting a 13% return.
- 33:57 – 35:37
Hiring Microsoft Alumni for AI Transition
John Belizaire emphasizes that the AI transition requires experienced personnel, not just power and land. Soluna hired Ryan Carver, Neil Robinson, and Mikhail Stark from Microsoft to bring over 55 years of hyperscale infrastructure expertise to its Dorothy 3 and Cadi 2 projects.
- 35:37 – 36:08
Construction Goals and Milestones
The company has built over 200 MW of Bitcoin data centers and aims to build its first 150-168 MW of AI data centers, with plans to double that capacity over the next few years.
- 36:08 – 37:36
Investor Watchlist for Next 6-12 Months
Investors should monitor lease announcements, closing of the Applied Digital joint venture, new projects reaching Notice to Proceed (NTP), and financing wins. The company is exploring three development options: vertical build, joint ventures, and capital-light licensing models.
- 37:36 – 38:54
Podcast Conclusion
Host Tim Gerdeman thanks John Belizaire for the detailed discussion, noting the company's strong start and growth potential. The episode ends with standard podcast sign-off and credits.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| Applied Digital Corp.APLD | equity | neutral | And so, and signed the JV with uh Applied Digital, a very experienced data center operator uh |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Soluna integrates into existing substations, reducing interconnection time to 20% of the time required for new connections (1-18 months vs 3-5 years).
Unidentified speaker (speaker_1) ·3:06
How to verify: Check Soluna's SEC filings (10-K/10-Q) for project timelines and interconnection status; compare with standard grid interconnection queue data. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSoluna has a project pipeline of almost 6.3 gigawatts and over 206 megawatts of operating projects.
Unidentified speaker (speaker_1) ·9:00
How to verify: Check Soluna's investor presentations or press releases for pipeline updates; verify operating capacity via company reports. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
The Briscoe Wind Farm is a 150 MW facility acquired in April.
Unidentified speaker (UNKNOWN) ·22:40
How to verify: Check SEC filings (8-K) for the acquisition announcement and asset details; verify wind farm capacity via public records or company disclosures. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSoluna targets a 13% return on projects.
Unidentified speaker (speaker_1) ·32:24
How to verify: Check Soluna's investor presentations or earnings calls for target return metrics. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
The Dorothy Data Center will create about 12 full-time jobs.
Unidentified speaker (speaker_1) ·21:16
How to verify: Check Soluna's project disclosures or local economic impact reports. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
CODI2 in Southeast Texas will create about 40 jobs.
Unidentified speaker (speaker_1) ·21:37
How to verify: Check Soluna's project disclosures or local economic impact reports. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Ryan Carver ran the entire construction program for AI at Microsoft, including the Fairwater campus in Wisconsin.
Unidentified speaker (speaker_1) ·34:16
How to verify: Check Microsoft's press releases or press kits regarding AI infrastructure construction; verify Ryan Carver's LinkedIn profile or Microsoft executive bios. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceNeil Robinson was head of construction for Microsoft in the Milwaukee region.
Unidentified speaker (speaker_1) ·34:36
How to verify: Check Microsoft's press releases or press kits regarding AI infrastructure construction; verify Neil Robinson's LinkedIn profile or Microsoft executive bios. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceMikhail Stark was responsible for enterprise planning and capabilities for Microsoft's AI and global cloud program.
Unidentified speaker (speaker_1) ·34:57
How to verify: Check Microsoft's press releases or press kits regarding AI infrastructure construction; verify Mikhail Stark's LinkedIn profile or Microsoft executive bios. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceThe three new hires have over 55 years of mission-critical and hyperscale infrastructure expertise combined.
Unidentified speaker (speaker_1) ·35:17
How to verify: Calculate the sum of years of experience listed on the LinkedIn profiles of Ryan Carver, Neil Robinson, and Mikhail Stark. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Soluna has successfully built over 200 MW of Bitcoin data centers.
Unidentified speaker (speaker_1) ·35:57
How to verify: Check Soluna Holdings (SLNH) SEC filings (e.g., 10-K, 10-Q, or 8-K) for project completion announcements or operational updates. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSoluna's goal is to build its first 150-168 MW of AI data centers.
Unidentified speaker (speaker_1) ·35:57
How to verify: Check Soluna Holdings (SLNH) SEC filings for capital expenditure plans or project development milestones. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSoluna has a joint venture with Applied Digital at Dorothy 3.
Unidentified speaker (speaker_1) ·36:08
How to verify: Check Soluna Holdings (SLNH) SEC filings for joint venture agreements or partnership announcements. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceSoluna is exploring a capital-light approach where tenants license power and land sites and handle infrastructure build-out.
Unidentified speaker (speaker_1) ·36:53
How to verify: Check Soluna Holdings (SLNH) SEC filings for business strategy updates or new business model announcements. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceRyan Carver previously worked at Microsoft.
Unidentified speaker (speaker_1) ·34:16
How to verify: Verify Ryan Carver's LinkedIn profile or Microsoft executive bios. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceNeil Robinson previously worked at Microsoft.
Unidentified speaker (speaker_1) ·34:36
How to verify: Verify Neil Robinson's LinkedIn profile or Microsoft executive bios. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Review Soluna Holdings' latest SEC filings (10-K, 10-Q) for financial health, debt levels, and related-party transactions.
- Verify the status and terms of the joint venture with Applied Digital, including capital contributions and governance structure.
- Check ERCOT's latest reports on curtailment rates and the 'firm power' classification criteria for data centers in Texas.
- Analyze the construction progress and milestones of the Briscoe Wind Farm and the Dorothy 3 project against the stated 150-168 MW AI capacity goal.
- Research the employment contracts and non-compete clauses of the newly hired Microsoft alumni to assess retention risk.
- Compare Soluna's interconnection timelines and costs against traditional data center developers to validate the 20% time reduction claim.
- Investigate the regulatory environment in Texas regarding data center permits and the 'firm power' bucket classification.
- Review competitor data center developers' strategies in the renewable energy space to assess market saturation and competitive advantages.
Full transcriptexpand
0:00Tim Gerdeman: Welcome to the WTR Small Cap Spotlight Podcast. I'm your host, Tim Gerdeman, co-founder of Water Tower Research. In today's episode, I'm joined by John Belazire, CEO of Soluna Holdings NASDAQ Ticker Symbol, SLNH.
0:18Unidentified speaker (UNKNOWN): Also
0:18Tim Gerdeman: joining is my WTR equity research colleague, Eric Goldstein. Soluna is a developer of green data centers for intent. intensive computing applications, including artificial intelligence and Bitcoin mining. The company's core idea is simple.
0:34Unidentified speaker (UNKNOWN): A
0:34Tim Gerdeman: meaningful share of the power produced by wind and solar farms goes to waste because the grid cannot always absorb it. Soluna builds data centers right at or next to those renewable power sources and turns that otherwise stranded energy into computing capacity. an approach the company calls renewable computing. Let me first start out by welcoming John and Eric to today's podcast.
0:59Unidentified speaker (UNKNOWN): Thanks
0:59Unidentified speaker (speaker_1): for having me on the show, uh Tim. Appreciate it. Yeah.
1:02Unidentified speaker (speaker_2): Good morning.
1:04Tim Gerdeman: John, to kick things off, many investors first came across your company as a Bitcoin data center company.
1:10Unidentified speaker (UNKNOWN): For
1:10Tim Gerdeman: a listener meeting Soluna Saluna for the first time today, or getting refreshed on the company's story, how would you succinctly describe Saluna's current business model and why does owning access to power put the company in a different position than a typical data center developer?
1:28Unidentified speaker (UNKNOWN): Well,
1:28Unidentified speaker (speaker_1): Tim, you you uh you described us well. We are a data center developer. The key difference between us and other data center developers is we go to uh locations around the country and hopefully the world, where uh renewable energy is wasted. So there are wind farms, solar, uh, hydro in some aspects that cannot monetize all of the power they produce. What we do differently is we co-locate our data centers. center. So we will go to those locations, uh, place the data center close by to that. wind farm and integrate directly into the wind farm substation, giving us access to that wasted energy, its non-wasted energy, and the grid, which together create a bundle of energy, if you will, that can support our entire data center needs for both our existing Bitcoin business and in the future for our AI business.
2:24Unidentified speaker (UNKNOWN): What
2:25Unidentified speaker (speaker_1): we'd like to say is while um power has become the central asset in powering, if you will, the future of innovation, we go to where the power already is, where that power asset is, and find that power hiding in plain sight. And we integrate that power resource into an AI infrastructure business. And um the approach that we take allows us to get access to that power at a much faster rate. Speed to power, not just power, is the product that customers are looking for these days. And using the Solunar approach by integrating behind the meter with that existing power generation facility, we are able to get access to power in uh 20% of the time. So what would take three to five years?
3:17Unidentified speaker (UNKNOWN): Uh,
3:17Unidentified speaker (speaker_1): we can do in a year to 18 months because of that approach. And the reason is because the grid sees our integration as a modification to an existing connection to the grid, as opposed to a brand new connection, which takes a whole bunch of study. That's very complicated. It doesn't like to make a lot of changes. Whereas we're enhancing a location where they have a problem. There's too much power being produced. They need load in that location. By integrating our data center, we solve a problem for the grid. We solve a problem for the wind farm. And we deliver a new asset to the future of intelligence.
3:59Tim Gerdeman: That's a fascinating differentiator. Uh With that, I'm going to turn it over to Eric to ask a few questions.
4:04Unidentified speaker (speaker_2): Great. Thank you, Tim. So John, thanks for joining us today. I know it's certainly been a pretty busy time for you. Your CODI 1 asset has been fully energized. You signed uh the applied digital joint venture for Dorothy 3. And yesterday you announced a new uh LOI for a tenant at CODI Coti 2. two and you paid off all your generate debt. So I guess before we get into the projects, um, maybe for the listeners, can you just tell us a bit about your background uh and the original strategy uh behind Soluna that centered around um the curtailed renewable energy?
4:46Unidentified speaker (speaker_1): Sure. Uh I'm I'm trained as a computer scientist, actually. I did my undergrad and grad. uh at Went to Cornell upstate New York. Started my career at Intel, where I was uh involved in helping to shape the future of personal computing. We were building a series of software technologies to help to drive more interconnection in the PC. So the convergence between telecommunications and uh personal computing
5:14Unidentified speaker (UNKNOWN): was.
5:15Unidentified speaker (speaker_1): the first uh, you know, call it first, first tile in the mosaic, if you will, for the future of technology that was about how computing was going to become a ubiquitous communication platform.
5:28Unidentified speaker (UNKNOWN): It
5:28Unidentified speaker (speaker_1): sounds funny to even hear myself say that because today we're so interconnected uh on a global basis to anything, content, information, intelligence, and and now. uh a social activity. And uh back then, this is uh over 25 years now, um that didn't exist. We didn't have uh the internet, we didn't have connectivity between computers. And so we were building the the first sort of networking platform for that. Uh after Intel, uh I went on to form a series of software technology companies that were built on the uh fast emerging big technology wave known as the internet and the whole dot com space. And we built a whole host of technologies to help businesses uh take advantage of that. And then later, uh as the next wave took place, which was uh you know digital uh platforms and uh social. interconnection and predictive analytics and so forth. We built software to help businesses take advantage of their big data, if you will, to make better business decisions. Most recently it was an insurance software platform that that uh I had built um to help large enterprises uh use data to make
6:43better underwriting decisions.
6:45Unidentified speaker (UNKNOWN): I
6:45Unidentified speaker (speaker_1): came to Saluna uh because one of my board members was uh
6:49Unidentified speaker (UNKNOWN): working
6:50Unidentified speaker (speaker_1): on this. gnarly problem they had where they were developing a uh wind farm in northern Africa that had this issue where they would not be able to monetize all the power. The grid was making its way down there. They had to they were planning to interconnect this this power plant, uh, but the the
7:11Unidentified speaker (UNKNOWN): the the
7:12Unidentified speaker (speaker_1): the market in that area couldn't consume all of the power they were. would generate. The size of this project was twice the size of Manhattan to give you a sense of scale. Could produce upwards of today, probably with current technology, close to a gigawatt of energy, but the market wasn't that big. And so what would you do with all the wasted power? And we found this uh you know intense computing application called Bitcoin that we can build there, designed uh a solution. to that problem by integrating a computing facility to to renewables, first of its kind. And um COVID happened and we had to bring that idea
7:52Unidentified speaker (UNKNOWN): to
7:52Unidentified speaker (speaker_1): the US because it'd be hard for us to travel to to the to the continent anymore.
7:57Unidentified speaker (UNKNOWN): We
7:57Unidentified speaker (speaker_1): started thinking to ourselves, does anyone else have this problem? And uh while we were locked up at home, we started calling grid operators, owners of big renewable energy power plant companies. Um, and that turned out to be easy to do because A, nobody else was traveling, so it was hard to get, it was easy to get people on the phone. And B, uh, the guys that were selling us the power equipment for our project in Africa knew all of the wind farms and solar farms that were dealing with this problem. So they connected us with these people.
8:28Unidentified speaker (UNKNOWN): And
8:28Unidentified speaker (speaker_1): then we reverse searched the locations, called the owners, and said, Hey, um, we have this problem. Do you have this problem? And they said, Yeah, we got it in a big way. It's a big problem for us.
8:40Unidentified speaker (UNKNOWN): Why?
8:40Unidentified speaker (speaker_1): What do you you guys have figured out how to solve this problem? Because we've tried everything. You know, there's nothing that can be done. And we explained to them the concept, and then before you knew it, a handful. of folks were like, come down here and help us. It was it was uh folks based in Texas. And that led to the platform that we have today, almost 6.3 gigawatts of of project pipeline, uh a number of some of the world top IPPs,
9:06Unidentified speaker (UNKNOWN): uh
9:07Unidentified speaker (speaker_1): over 206 megawatts of operating projects, um three phases of Dorothy and the third first uh third phase um becoming AI and now being an owner of renewable energy generation projects ourselves. And so we've come a long way from this you know very simple idea of convergence. Um we believe that renewables and computing are quickly becoming one asset, and AI is accelerating that. integration. And uh in the future, we think that will result in more power, more renewable power being built around the globe.
9:48Unidentified speaker (speaker_2): That's great. Thank you. That's a great answer. So, just for listeners, uh, just quickly who are not steeped in power markets, how does building behind the meter at a renewable site work? Um, you know, particularly like in Texas, where you guys are mainly located.
10:05Unidentified speaker (UNKNOWN): So
10:05Unidentified speaker (speaker_1): how it works is so uh imagine you're you're standing in in some giant uh ranch, and above above you and all around you are these giant wind turbines that are spread, you know, sprawling across thousands of acres, and it's a windy day. And um you notice something peculiar about half, let's say there's a hundred turbines there,
10:31Unidentified speaker (UNKNOWN): about
10:31Unidentified speaker (speaker_1): half of them are spinning and half are not, even though clearly there's wind blowing across the ones that are not spinning. Why is that? Are those turbines broken?
10:40Unidentified speaker (UNKNOWN): Uh
10:41Unidentified speaker (speaker_1): it turns out that
10:42Unidentified speaker (UNKNOWN): the
10:43Unidentified speaker (speaker_1): owner of the wind farm uh has been told by Urcot ERCOT that they have to turn off half of the plant because uh today, or they told them yesterday actually, that they wouldn't be able to take all of their power. And so they signal their systems to turn off, you know, turbines uh, you know, one through 49, right? And the rest are on uh producing power. That goes through the substation. Those electrons flow through the substation onto the grid and supply. let's say this wind farm is a hundred megawatt wind farm, supplies 50 megawatts to the ERCOT grid, those electrons flow across the high voltage lines and meet their power users, if you will, down in some of the big cities and so forth, if you will. That's before we arrive. Then take that same movie and imagine this. little truck pulls up. Yeah, it's got a it's got a circle circle delta logo on the side. A company called Soluna. Guy pops out with a black shirt, circle, circle delta on it, and says, Hi,
11:51Unidentified speaker (UNKNOWN): I
11:51Unidentified speaker (speaker_1): notice you you have a little curtailment here. That's what you call that scenario where some of the farm is producing and some of it's not. And the owner says, Yeah, that's right. Um what of it? We say, Well, we have a great solution. Um, if you allow us to get some of this land, we'll build a data center and we'll integrate to the farm to the farm. And even though your signal to curtail, you'll still be able to turn on and we'll use that
12:22Unidentified speaker (UNKNOWN): uh
12:23Unidentified speaker (speaker_1): unused power. So those other 50 turbines will now turn on. So fast forward about six months to a year, the facility is now running. Uh, it's connected to the wind farm. And most days, when it's windy and you know, the turbines aren't broken, most of the turbines are spinning now, um, depending on the size of the data center we built. Let's say we built a 50-megawatt 50 megawatt data center. then uh
12:53Unidentified speaker (UNKNOWN): you
12:53Unidentified speaker (speaker_1): know what would have been half the facility not spinning, the entire facility is spinning now. And so they doubled their revenue, right? Because we showed up and offered them, offered to buy that power from them. Now, if we make the data center 100 megawatts, for example, then we'll use the 50 megawatts that would otherwise. be wasted. And we'll also use the other 50 megawatts that was going into the grid. And now the load is local to that rather than it has to travel all the way down to Dallas, let's say, for example, if we're in West Texas, serving load that's local to that That wind farm now. And so now it has a hundred megawatts of its power being generated. uh, lifting up the revenues. If the wind farm has any kind of tax credits, they'll get those tax credits now.
13:46Unidentified speaker (UNKNOWN): And
13:46Unidentified speaker (speaker_1): as a result, be able to pay back all their loans and and make a healthy profit in the process. So it's a very simple idea, you know. Um we may we've even taken it one step further instead of the pickup truck showing up and saying, Hey, let us build a data center and you know, we'll connect into your wind farm. We've actually shown up with a bigger truck and say, hey, on the side of the truck, and it says, we'll buy your wind farm. Um, your buy your trouble wind farm. Um and uh
14:15Unidentified speaker (UNKNOWN): we
14:16Unidentified speaker (speaker_1): turned that wind farm into a power generation facility facility for an AI campus. That's exactly what we're doing at our Dorothy uh project. Um we started out with just buying. the power and curtailing it and preventing curtailment and building up and scaling up to 100 megawatt uh Bitcoin data center. Now we own the 150 megawatt wind farm. We acquired it. And we're moving that power across the street to power uh a multi-hundred megawatt AI campus. And so that wind farm will become a lot more valuable after it is the primary source for this AI campus. And we want to do that again and again to really drive next generation economics to these assets and allow folks to build more of
15:07Unidentified speaker (UNKNOWN): them.
15:08Tim Gerdeman: Eric,
15:08Unidentified speaker (UNKNOWN): I
15:08Tim Gerdeman: have to comment quickly um that I appreciate John's very thorough answer because for years and years I've driven for more from Orange County, California over to Pullm Desert. And I always scratch my head saying, How in the world can this be an economical model when maybe one third of the turbines are actually turned
15:25Unidentified speaker (UNKNOWN): on? Yeah,
15:26Unidentified speaker (speaker_1): a lot of people don't know this, but yeah, it's it's uh a lot of these electrons never actually make it to the grid.
15:32Unidentified speaker (UNKNOWN): And it's for
15:33Unidentified speaker (speaker_1): that reason.
15:34Unidentified speaker (speaker_2): Makes it make makes a lot of sense. Uh and I had the same curiosity as as Tim. did as well. Um, so uh can we just maybe have a brief discussion on uh just the midterm elections and data centers, AI have you know obviously become sort of a political hot potato. Um so I I guess what are you seeing in Texas and then and how are you guys dealing with the changing landscape if you see it changing at all?
16:04Unidentified speaker (speaker_1): So uh there's a lot of talk of change, but there's not much change yet. Uh prior to the midterm, uh the sitting governor began implementing a series of executive orders to uh essentially uh slow down the data center industry, put a bit of a pause on it. Um, that the those included a an audit of uh power uh requests. So PerCOT was seeing over 400 gigawatts of power requests for a market that is upwards of 80 to 90 gigawatts in terms of its load. So they were scratching their heads like,
16:45Unidentified speaker (UNKNOWN): this
16:45Unidentified speaker (speaker_1): this is impossible. All of this, all of these requests can't be real. So the governor wanted to understand how many of these are real, how many of them that them have the financing. What are they going to build exactly? And what is that going to do to the communities, right? And so he he requested an audit, which is still underway, is scheduled to complete uh early December. And the second thing he did was uh pause. permits, um, the permitting system for certain projects um on the environmental side, water and uh um air permits, things like that that you would use for for generators or for um gas fired power generation, which was a lot of stuff that was underway. And so uh what that spawn spawned was a series of of uh activities. ERCOT had to um classify some of these 400 megawatts, if you will, into different buckets. Uh and those buckets determine how firm your power was. And in Soluna's case, we fell into one of the uh firm power buckets because of our
17:58approach and the fact that
18:00Unidentified speaker (UNKNOWN): uh
18:00Unidentified speaker (speaker_1): a number of our projects were already energized. And if you were still studying as to whether you would connect or build a data center, then you would be on one of these um less firm buckets, and that increases the risk of your project.
18:18Unidentified speaker (UNKNOWN): And
18:19Unidentified speaker (speaker_1): then on the permit side, um, essentially permitting was paused. So if you had a project that was under construction and you didn't have a permit that you needed to either start construction, complete construction, or have your customer move into the building, let's say, you were going to be paused. Thankfully, the two projects that we have there are under construction. And the permits that we need allow you to at least start some construction until you get to the phase where you need to get the permit. So we have some time. And as we speak, as we're recording here, we're starting to see a lot of those blockers removed, if you will. The permitting process is starting to reopen. Some of the key agencies that were told to not take submissions are now taking submissions again. Obviously, they're going to be quite backlogged now, so things are going to take a little longer. longer.
19:16Unidentified speaker (UNKNOWN): Um,
19:17Unidentified speaker (speaker_1): and the audit process, as I said, will come to an end. Um, and everybody has, you know, submitted their responses to the RFI that was sent out and so forth. So that's the current administration.
19:28Unidentified speaker (UNKNOWN): The
19:29Unidentified speaker (speaker_1): next administration um is signaling two things. One is um they would like to see more. uh conscious intentional behavior by data center developers to um be good actors in the state. They'd like to see processes that uh weed out speculation, right? That um
19:55Unidentified speaker (UNKNOWN): that
19:55Unidentified speaker (speaker_1): was part of the problem that uh Governor Abbott was trying to address.
20:00Unidentified speaker (UNKNOWN): Um
20:00Unidentified speaker (speaker_1): and then the third thing is uh taking a step back and looking at how much the data center economy is driving the state's economy and making a decision as to how much we want to curtail that now that the mandate is to grow grow economies and create jobs. And so
20:23Unidentified speaker (UNKNOWN): when
20:23Unidentified speaker (speaker_1): we look at it from a solution perspective, what we enjoy is the fact that a lot of the concerns that the current governor has and the governor-elect has had with this industry, we already address. A, we invest in infrastructure that integrates with existing plants that have wasted energy. So we're not raising the cost of energy. In fact, we don't even use the same electric. electrons that a typical household in the communities we're we're uh we're going to be based in are using. We use uh we build data centers that use almost no water. Uh they use a closed loop system where you inject a special magic water into the building and it just circulates forever and it's air cooled. Um we uh create jobs. Our uh Dorothy Data Center, we think, will create about um 12 uh full time jobs. It it generates you know thousands of construction jobs, but that's temporary, and then it um resolves to about uh 12 direct full time jobs, and then another you know uh uh tens of uh
21:37jobs on the sort of indirect. mode.
21:40Unidentified speaker (UNKNOWN): And
21:40Unidentified speaker (speaker_1): in CODI2, our much larger data center down in Southeast Texas will be about 40 jobs. So um we are creating uh jobs, we're creating new industry in these these areas, and we're doing it in a sustainable way. And so from our perspective if uh Saluna represents a a uh a blueprint um a a bright spot for how you should do data center development um in Texas and beyond.
22:08Unidentified speaker (UNKNOWN): M
22:08Unidentified speaker (speaker_2): makes perfect sense. So in April um you bought the 150 megawatt uh Briscoe wind farm and the remaining interests in Darthi one A and uh Darthi one B. So just why was owning the power plant itself important? Does that help you uh market the project to prospective tenants? Uh, and do you envision you know potentially buying other wind farms or other you know renewable generating assets in the future?
22:40Unidentified speaker (UNKNOWN): So
22:41Unidentified speaker (speaker_1): the reason we bought the wind farm is because um energy Dorothy's sovereignty is becoming increasingly important um in this market, as I made mention of to earlier, power and compute are quickly becoming one and the same thing. And so it means that these campuses kind of need to own those assets to have better control of them, but more importantly, better control to the speed of conversion of those electrons to compute.
23:10Unidentified speaker (UNKNOWN): So
23:10Unidentified speaker (speaker_1): in our current construct, prior to buying Dorothy, we would. create bilateral agreements with the owners of these power plants, like I mentioned.
23:21Unidentified speaker (UNKNOWN): They
23:21Unidentified speaker (speaker_1): sell us that power. They give us the ability to integrate to their substation in return and get access to grid power. We got to put in all these structures. And that works, the contractual relationship. But the speed of that, or at least part of the speed of that, is outside of our control. How fast. the IPP moves, whether their partners are interested in doing this deal or not. Other uh uh concerns by different uh entities in the capital stack, et cetera, slows things down to some extent. By owning the asset, we now have total control of both sides of the picture. Uh we have a dedicated, focused uh strategy for the conversion of the power to uh supporting the AI campus. And so it allows us to uh do the conversion much faster. And so speed of convert converting the Dorothy project was uh uh uh a key importance to us to give you a sense of that speed. We completed the acquisition of uh Briscoe in
24:24April, April 1st, I believe. we we made the announcement. We're now sitting in November. We are already uh constructing the first phase of our AI data center at that campus. And so, and signed the JV with uh Applied Digital, a very experienced data center operator uh who's partnering with us on the site, and it's gonna going to be um uh hosted. hosting a uh hyperscale customer in there. So that's less than a year from acquisition to conversion.
25:00Unidentified speaker (UNKNOWN): It
25:01Unidentified speaker (speaker_1): would have been very difficult for us to do that um if we didn't own the asset. And we've already invested in, as you saw in our financials, uh upgrading and improving the technology of the wind farm so we can produce and work better. I was there last week. and the entire wind farm, every single turbine was spinning. It was a windy day. Every single turbine was spinning. It was looking great. And it felt great. I went from, wow, this is, you know, when we first uh started talking to this, uh to the IPP that owned it. You know, we were standing on just like empty dirt and we're thinking to ourselves, you know, we're looking at this wind farm. Like I said in my story, you know, half of it wasn't spinning, and we went to visit. And, you know, over the course of the last few years, we've
25:43Unidentified speaker (UNKNOWN): did
25:43Unidentified speaker (speaker_1): the first phase, the second phase, doubled the size of the plant and kept adding, you know, more value to the wind farm. And um, you know, during the acquisition talks, you know, it was clear that we we were driving most of the value for the for the asset. It just made sense for us to own it. And now uh less than a year. after owning it, we're already on our way to significantly increase the value of the asset and uh turn those electrons into their next best, most high economic use. And that completes the, you know, once we once we get that and that building is finished, that completes the entire story. We've now proven that, uh, or we will have proven that. you can now turn these types of assets,
26:31Unidentified speaker (UNKNOWN): these
26:31Unidentified speaker (speaker_1): highly curtailed wind farms, into digital AI infrastructure, into into um factories. And what that will trigger, we hope, will be that the renewable energy industry will become computing companies. They will go build renewable power all around the world to power computing. compute and the grids that they're going to be connecting to, which will result in more green energy being on the global grid.
27:01Unidentified speaker (speaker_2): Okay. Uh in October, um, you just you just mentioned Applied Digital. So in October, you you uh signed an agreement with Applied Digital for uh 50 megawatt uh uh uh uh critical IT firm. first phase at Darthy3. And you're gonna own 49%, they'll own 51%. So um why partner rather than go it alone? And uh what's left to do before the end of the year just to turn that into uh an official deal
27:33Unidentified speaker (speaker_1): so um the reason to the partner is uh capital cost of capital These builds are multiple times the, you know, there isn't there isn't even a comparison when you compare what it costs to build a Bitcoin data center, which is in the you know hundreds of thousands of uh dollars per megawatt to
27:58Unidentified speaker (UNKNOWN): an
27:59Unidentified speaker (speaker_1): AI data center, which is in the tens to teens millions. of dollars per megawatt, you're talking about a tremendous amount of capital that needs to be deployed. Most of that is debt. But these debt providers want to back experienced operators. They want to provide, we want to get access to that debt at a reasonable cost. And so by partnering and bringing a partner that has lots of experience. access to these capital partner relationships, it allows us to essentially half the cost from an equity perspective, right? That we would need to deploy. And that puts less pressure for us to go, you know, d dilute our shareholders to go raise that capital. So by having, you know, two checks at the table, we can build more, uh, build it faster. uh, build it at a lower cost.
28:54Unidentified speaker (speaker_2): Okay. So as you said, just for for listeners, um, you know, these are very capital-intensive projects. So you guys have shown that a a 100 megawatt campus um will probably cost in the range of 1.2 to 1.3 point three billion dollars. And about 80% eighty percent at the project level would come from debt. and you know roughly two hundred and fifty million uh would would have to come from Solana in in terms of equity so
29:23Unidentified speaker (UNKNOWN): can
29:23Unidentified speaker (speaker_2): you
29:23Unidentified speaker (UNKNOWN): just
29:24Unidentified speaker (speaker_2): yeah so maybe you can just talk about that a little bit um how how would you expect um to to finance you know a project that size
29:34Unidentified speaker (speaker_1): uh so the uh the the the capital capital stack stat um um is uh is is not something that we're we're we're inventing, right? Um data centers have been uh built and have been uh expanding in the world for the last 20 plus years, right, since the internet. And so there's a well-established uh financing model for these assets. It's about anywhere from 70 to 80 percent debt, meaning project. level debt. So debt that attaches to the data center itself as an asset. It's its primary collateral. It's attaching to the cash flows, the net operating income, which is the cash flow that's going to come from the lease revenue from the tenant, and um minus any cost uh associated with running the building. is gonna give you a set of free cash flows. That's gonna pay off the debt,
30:32Unidentified speaker (UNKNOWN): uh
30:32Unidentified speaker (speaker_1): portion of the debt payments uh or amortization, and then you have sort of your
30:38Unidentified speaker (UNKNOWN): you know
30:38Unidentified speaker (speaker_1): net amount of capital that's that's available for for for equity, etc. And uh, you know, these these data centers at a 1.3 you
30:46Unidentified speaker (UNKNOWN): know
30:47Unidentified speaker (speaker_1): billion dollar build, you could uh 100 megawatts, you can easily see.
30:51Unidentified speaker (UNKNOWN): you know,
30:52Unidentified speaker (speaker_1): call it 150 million dollars, right? And NOI. So that gives you a sense of the top line. There's plenty of cash flows to cover a good deal of debt.
31:05Unidentified speaker (UNKNOWN): The
31:05Unidentified speaker (speaker_1): result is that you have about um
31:08Unidentified speaker (UNKNOWN): you
31:08Unidentified speaker (speaker_1): know 20% to a third of that, which is equity. And so that is the equity that comes from ourselves at a parent level or a project level partner, for example, through a JV, that finances that portion. And that's your full capital stat. So a combination of um
31:33Unidentified speaker (UNKNOWN): You know,
31:33Unidentified speaker (speaker_1): project level equity and uh and project level debt. And we structure these deals to generate a a yield above the cost of capital. So that debt has a cost to it. We don't give you money for free. And so you want to structure the deal such that there's a healthy margin above that cost of debt to have a healthy amount of cash flows to equity, but also uh reduce the risk of not covering that debt. right? And um we try to structure it in a way that we have that margin and then we back into the lease rate, the minimum lease rate that we have to hit in order to finance the project. So I like to say it's just, you know, with these projects, it's just math, right? It all kind of works out to a point where you've got to generate a target return. You know, um, I think we've said publicly in our illustration about like a 13% um target return. That's creeping up these days because cost of capital is going up with uh it's happening in the market as we're as we're recording here at AI trade and uh high-yo bonds kind of you
32:44know closing. So um we're seeing that that creep up and that and therefore um leases and lease rates are are also creeping up. And so um we're structuring everything basically around a formula such that we have the cost plus some margin uh gives us enough comfort that we can we can finance the project.
33:08Unidentified speaker (speaker_2): Okay. So um you know as we look forward, I guess over the next um, you know, one to two years. there's certainly a lot of uh growth on the come based on all these projects for your company. Uh it seems like you've uh you've been solidifying your team. You've brought in, you know, someone from Microsoft. I think you have a new board member from Google. So maybe, maybe you could talk about your team just briefly and what gives you the confidence you'll be able to execute on on all the things you have in front of you. And then maybe for investors who are listening, um, maybe kind of give them um you know things that they should look for over the next six to twelve months um for the from the company.
33:53Unidentified speaker (UNKNOWN): Uh
33:54Unidentified speaker (speaker_1): absolutely so um
33:56Unidentified speaker (UNKNOWN): you know
33:57Unidentified speaker (speaker_1): we're we're going through an AI transition here um and that transition is more than just power and land uh converting those to to to data center assets. You also need uh people who uh who have built this before. And um, you know, I like to say in my entrepreneurial experience, um, mistakes are okay, but only if they're original mistakes. Non-original mistakes can be avoided by bringing experience in. Uh so we hire hired Ryan Carver as chief development officer. He ran the entire construction program for AI at Microsoft, including one of their largest campuses called Fairwater in Wisconsin. Neil Robinson was head of construction and delivered uh a number of campuses uh for or Microsoft in uh the Milwaukee region. as well. And uh uh Mikhail Stark, who's head of planning, and uh
34:57uh she comes from Microsoft also. She was responsible for uh enterprise planning and capabilities for uh not just AI, but their entire uh global cloud program. Combined, I think these three folks have over 55 years of mission critical and uh how. hyperscale infrastructure expertise. And so now they're they're bringing that expertise to renewable-powered AI data center campuses at uh Dorothy 3 and Cadi 2. So, you know, in our mind, we like to say power is the asset. And bringing these people on makes it clear that uh people uh make it real. And um you're your question about what to expect in the coming months, well, these folks and a host of more people we're going to hire are going to help us to bring these projects to life. We have successfully built over 200 megawatts of these Bitcoin data centers. Now
35:57our goal is to uh build our first 150, 168 megawatts of AI data centers and then keep doubling that over the next few years. You should uh be looking for uh lease announcements, uh closing announcements around the closing of our applied uh JV. You should also uh be on the lookout for new projects that uh we're going to be uh reaching NTP and starting the process of marketing and developing. There's a series of material we put out that show those future projects and their stages. And those will be developed across three different options. One might be going vertical, like we're doing now for Gotti 2, a joint venture, like we're doing in DoorT3, or there's a new emerging approach where tenants or uh larger operators are licensing power and land sites, and then they do all of the infrastructure build out. So that's a capital light approach that allows us to monetize our assets without having the to to build the the whole campus, if you will. So that's something we're we're exploring. We haven't
37:13announced anything there yet. but it's is certainly something we're considering. But there have been announcements of those types of deals in the industry. So that points to where things are heading in the future.
37:24Unidentified speaker (UNKNOWN): So
37:24Unidentified speaker (speaker_1): I think what you should be looking for is um uh leases, construction starts, and uh financing wins. So getting the capital in the door to to build these out. up.
37:36Unidentified speaker (speaker_2): Okay. That's great, John. Thank you very much. It's certainly a very exciting time for the company. You have a lot going on. And you know, and if you execute on these projects, which it seems like you're off to a great start, the business has uh unbelievable growth potential and something that uh listeners and shareholders um uh should be looking forward to. So we look forward to to tracking all the mile. milestones and uh we we'd love to have you back.
38:06Unidentified speaker (speaker_1): Thank you. I'll be back and uh give you guys updates as we make progress.
38:10Unidentified speaker (UNKNOWN): Well,
38:10Tim Gerdeman: thanks, Eric, for those thoughtful questions. John, thank you so much for uh walking us through in great, great detail. Um, Saluna, great, very, very interesting business model. Feels like you're at the right place at the right time. So best wishes.
38:22Unidentified speaker (speaker_1): Thanks, Dan. Appreciate the time.
38:26Unidentified speaker (UNKNOWN): Thank you
38:26Tim Gerdeman: for listening. and don't forget to subscribe as well as visiting WW Watertower Research.com to stay up to speed on the company's small cap written research reports, podcasts, fireside chats, industry-specific symposiums, and conference schedules. We will see you next time for another edition of the WTR Small Cap Spotlight Podcast. Finally, a special thanks to the producer and editor of the podcast. Krista Fitzpatrick.