Soluna Holdings, Inc. (SLNH) Earnings Call Transcript
August 13, 2026
Earnings Call Speaker Segments
Good afternoon, everyone, and thanks for being on Soluna's call for the second quarter of 2026 to discuss our earnings. You can find our earnings release along with the presentation that goes with it over in the Investor Relations area of solunacomputing.com, and there's a webcast of this call running alongside those slides. Joining me are John Belizaire, our Chief Executive Officer; Ryan Carver, our Chief Development Officer; and Michael Picchi, our Chief Financial Officer. Ahead of management's prepared comments, we want to remind all of you that certain things we say today might count as forward-looking statements under securities laws and carry a range of risks and uncertainties. Because of that, we caution that many factors exist, a great number of them outside our control, that might lead actual results and events to come out materially different from what those forward-looking statements describe. If you want more detail on the risks, uncertainties, and assumptions tied to our forward-looking statements, please review the disclosures in our earnings release and the public filings we've made with the Securities and Exchange Commission; we disclaim any obligation or undertaking to update forward-looking statements to reflect circumstances or events that occur after the date the forward-looking statements are made, except as required by law; we will also discuss non-GAAP financial metrics and encourage you to read our disclosures and the reconciliation tables to applicable GAAP measures in our earnings release carefully as you consider these metrics. For thorough disclosures and descriptions of our business, plus uncertainties and other variable circumstances, including, but not limited to, risks and uncertainties identified under the caption Risk Factors in our annual report on Form 10-K for the year ended December 31, 2025, and our other SEC filings, we point you to our filings with the Securities and Exchange Commission. One thing to flag before we start: starting this quarter, pass-through electricity costs are presented on a gross basis rather than the net basis we used previously. Reported revenue and reported cost of revenue both rise by the identical amount under that change, so gross profit, operating loss or net loss are unaffected. Each slide that's in effect carries a footnote, and Mike will take you through how it works. With that said, I'll hand things to John.
Appreciate that, Mike Tu. Hi everyone, and thanks for joining Soluna's second quarter results and business update; this marks our inaugural quarterly earnings webcast, and we're happy to have each of you with us. I'll begin with the thesis, since everything else we cover today rests upon it. In the AI era, power is the main constraint — not chips, not capital — power, and how fast you're able to get to it. Soluna has locked in long-term behind-the-meter access to power at scale, and we turn that access into contracted data center flows and data center cash flows by building right on the site of renewable generation with stranded power, skipping long grid queues. There are four figures that capture our current position. Our top line climbed 145% compared with the same quarter last year, marking a fifth consecutive quarter of quarter-over-quarter expansion. We manage 192 megawatts of capacity — an operating base that reliably produces cash — and by this summer we anticipate that figure reaching 206 megawatts. Across 2 AI campuses, Kati 2 and Dorothy 3, we have more than 650 megawatts in development, with both sites on schedule for locking in leases and construction. Meanwhile, our pipeline for renewable power expanded 47% over the year, surpassing 6.3 gigawatts — the hard-to-find input for AI, all procured behind the meter. Let me now turn to the second and third items, since during this quarter those two came together. Our philosophy is that power represents the true asset while compute trails behind it — that's what we call the Soluna Way. At Project Dorothy, we carried this approach through completely during Q2. We finalized our purchase of the 150-megawatt Briscoe Wind Farm on April 1. Then on April 15, we bought out Spring Lane Capital's stake in Dorothy 1A. Following that, on May 19, we picked up Navitas' share of Dorothy 1B. As a result, Soluna holds both the generation side and the computing side for the entire 50 megawatts that make up Project Dorothy 1. Right now, the whole sector is in a scramble for electricity—locking in power purchase agreements, holding spots in interconnection lines, and haggling with utilities over capacity that won't arrive until 2029. That's why we went ahead and bought the wind farm. So when a hyperscaler or a neocloud player is doing due diligence on one of our locations, they boil everything down to two questions. First, how soon can power be delivered to me? Second, what will my energy cost be? For both, we point to an asset that we fully own. That approach now yields five separate revenue streams, and we're intentionally changing the balance among them. Our biggest business at present is hosting for Bitcoin miners: we construct and run the data center, the mining firms with solid credit bring in their equipment, and that's what fuels our top line. When it comes to proprietary Bitcoin mining, that's handled by our very own fleet; we sell each day and keep nothing in a treasury. Through grid ancillary services, we get paid for being flexible, interruptible load, and compensation comes our way for serving as behind-the-meter flexible load on behalf of the grid operator. Soluna Wind joins us as the new piece added this quarter, bringing on-site wind generation at Briscoe together with an ERCOT substation and grid interconnection. Then, naturally, there's AI high-performance computing, colocation and hosting for businesses requiring AI-ready capacity — that's the direction this company is headed. Ryan stepped in specifically to build that area. Everything we've pursued this year is built on the 4 priorities. First up, create AI. Move Kati 2 and Dorothy 3 forward to shovel-ready and tenant-ready status, and assemble a pipeline of AI-ready campuses built for swift deployment. Following that is making our projects as efficient as possible. Our attention is on powering up and ramping Kati 1 and generating profitability throughout the operating fleet via uptime, operational efficiency and disciplined cost management. Then comes capital formation; our plan is to finance pipeline growth and AI development by way of project level financing and strategic capital partnerships, all while keeping balance sheet flexibility. And last, our focus is on expanding our pipeline. Like I said earlier, this quarter we reached 6.3 gigawatts, and more than 300 megawatts of that increase came from expanded term sheets at 4 sites that are already in our portfolio. Since capital formation is what people ask me about the most, I want to talk about it; during the second quarter we raised $159.4 million, and then after the quarter closed we brought in an additional $23.6 million through our ATM program by issuing about 18.8 million shares. So here's what that accomplished. Full ownership of Project Dorothy 1, which is a 150-megawatt wind farm, the Kati 2 joint venture, plus land for Dorothy 3, and a clean capital structure; this quarter we fully retired the Series B and paid out the accumulated dividends. Every dollar of that went into assets that remain on our balance sheet as of today. When we look forward, the big AI construction projects are set up so that project-level debt provides most of the funding, with the data center serving as collateral and underwriting based on contracted tenant cash flows. Mike Picchi is going to take you through how that structure works. Now, let's get into the quarter. On the business front, there are four items. During the most recent reconstitution, we were added to the Russell 3000 and Russell 2000 indices, and over the past few weeks, new sell-side analysts have begun covering Soluna. Together with the more formal quarterly updates you're hearing today, these reflect a deliberate push to make Soluna more visible and more accessible to institutional investors. As I noted earlier, on April 1 we completed the purchase of Briscoe Wind Farm for $53 million, which marks the first time we've owned a renewable generation asset outright. We brought 100% of Project Dorothy 1A and 1B into consolidation, which bolsters our progress toward Dorothy 3, our new AI campus. When it comes to projects, our teams came through at all 4 campuses. Kati 2 finalized a joint venture with Metrobloks, wrapped up design development and got a tenant letter of intent signed. Dorothy 3 locked in a definitive land purchase agreement and moved utility coordination forward. Kati 1 finished substantial construction, while Dorothy 1A and Sophie preserved capacity through a summer curtailment period that was heavy. Ryan will walk you through the AI project highlights in detail in a moment. Let me turn to the pipeline now, since that's really the foundation of what we have. When we look at August 1, we're sitting at roughly 6.3 gigawatts across the whole pipeline. Earlier in the year that figure stood at 4.3 gigawatts, so it has grown, and I'd like to walk you through the way it breaks down. At present, 192 megawatts has reached operation, is generating, and is producing revenue for us. Construction is underway on a further 14 megawatts, which represents the final phase of Kati 1. We expect that once summer wraps up, we'll be at 206 megawatts operating. Sitting in planning and development is about 1.6 gigawatts, and within that bucket we're making headway on PPA negotiations, ERCOT planning, AI feasibility studies, and land acquisition. On top of that, roughly 4.5 gigawatts sit in assessment together with our power partners. There's another piece to how this model works: thanks to the behind-the-meter configuration we use, our data centers remain flexible. Electricity is able to flow either from the renewable plant or from the grid, and we can push ancillary services back onto the grid. That flexibility explains why we're able to interconnect so quickly. Clustering also figures into what we're planning. We intend to place several generation assets close to a single data center site, and that's what lets a footprint capable of holding just part of the capacity become a 300-megawatt campus. This is the approach we're using at Kati 2 and Dorothy 3, and it's going to appear at our remaining sites as well. Before I get into the road map, I'd like to speak to the recent announcements out of Texas.
Back on August 3, Governor Abbott instructed the PUCT along with ERCOT to review all data centers sitting in ERCOT's interconnection queue prior to greenlighting any new projects. This came about in reaction to pending requests totaling somewhere around 474 gigawatts, with data centers making up about 90% of that. The directive applies to newly studied loads waiting in the interconnection queue. I'd like to walk through 4 points regarding where Soluna stands. To start, the exposure we face directly is fairly small. Since the audit is aimed at new studied load, close to 146 megawatts of our Texas capacity has already reached energized status, and each of Dorothy 3 and Kati 2 is being built from energized interconnections next door. Next, the way we operate matches exactly what the state wants to see: we avoid expensive transmission upgrades, we offer load that is flexible and interruptible, we bring in fresh wind and solar, and we keep water use to a minimum through our design. We pay for our own electrical infrastructure. Third, we view this as a favorable wind for companies holding live capacity. Because the queue is frozen while the audit proceeds and ERCOT pushed back its August 7 dispatch, capacity that is already energized becomes more valuable. And fourth, we embrace this review. A strict and uniform review process is something we back, and we're giving our complete cooperation to the PUCT and ERCOT. So, to wrap up, let me cover what lies ahead for the rest of the year. Our attention is on finishing Kati 1, specifically the last 14 megawatts. At Kati 2, we intend to shift from design development into construction documents. We're pushing the Dorothy 3 development forward and starting to market it to prospective tenants. We anticipate fresh Bitcoin hosting announcements happening at Kati 1. Beyond that, we're exploring additional power purchase agreements now in progress involving Rosa, Hedy, Ellen, Annie and other new projects. Those are the milestones, and we'll give you an update against them next quarter. All right. Before passing it along, everything discussed so far — the wind farm, the buyouts, the capital — is there for a single purpose: turning our power position into contracted AI megawatts, which leads me to the newest addition to our leadership team. On July 16, we named Ryan Carver our Chief Development Officer. Ryan comes to us from Microsoft, where his most recent role was Senior Director of AI Construction and Site Development, overseeing a construction P&L in the tens of billions of dollars spanning the company's AI data center development. Very few individuals have taken hyperscale AI campuses from start to finish, and Ryan is definitely among them. His mandate here is intentionally wide, spanning site selection and development through engineering, construction, and operations, all within one organization that reports to me. Ryan, welcome to Soluna, and the floor is yours.
Thank you, John. It's a genuine pleasure to be part of Soluna. Hello, everyone, and good afternoon. Having only been here 4 weeks, I'll skip a lengthy resume and focus my remarks on the projects. As John said, my time at Microsoft spanned over 10 years, and my final role there was Senior Director of AI Construction and Site Development. In practice, that role involved shepherding AI campuses from raw land all the way to top-tier functioning data centers: securing power, obtaining permits, engineering, building, commissioning, and transitioning to operations. The standout initiative in my portfolio goes by the name Fairwater, located in Mount Pleasant, Wisconsin. My expertise lies in delivering big, complex infrastructure within schedule and within budget. So why did I join Soluna? Over recent years, I've observed this sector hitting the identical obstacle. You can purchase chips, but power requires waiting, and buildings must be constructed. Nearly every solution I encountered boiled down to waiting your turn — your place in the queue, the negotiation with the utility, capacity arriving many years behind what might otherwise be possible. The approach Soluna takes is the strongest I've come across: situate the data center behind the meter, placing facilities alongside sustainable generation sites that already exist, are underused, and are already wasting energy. That route is shorter and verifiably more sustainable while delivering the very same megawatt. My responsibilities encompass the AI campuses we are getting ready to build. Before that, though, I want to walk you through the campuses Soluna operates right now. Kati 1 is our 83-megawatt campus located in Willacy County, and it's essentially finished. K1A, which is what we call the Galaxy Digital portion, accounts for 48 megawatts and remained stable throughout the summer 4CP constraint window while also satisfying its ERS program obligations. K1B represents the build-out. Phase 1 consists of 12 megawatts of Cormint containers, and Phase 2 adds 9 megawatts of Soluna-designed data centers. Right now, both are energized and running. Phase 3, the last 14 megawatts, is under construction and progressing ahead of schedule. This year, Kati 1 generated its first positive gross profit. A site transitions from consuming capital to producing cash precisely one time, and that transition happened for Kati 1 this quarter; now shifting to the Briscoe Wind Farm that John referenced earlier. Briscoe is a 150-megawatt wind farm situated in West Texas. We bought it on April 1 for $53 million. It marks Soluna's first instance of directly owning a generation asset, and it has now been folded into our operations; along with the acquisition, we took on a maintenance backlog, and we chose very deliberately to address it right away. That effort wrapped up in Q3. In addition to the turbines, we closed out our Q2 renewable energy credit sale and carried out considerable community engagement with county officials and nearby landowners, which is important for everything we plan to construct in that area. Now turning to Kati 2. Kati 2 is positioned directly across the street from Kati 1. Once fully built, it should exceed 350 megawatts. Phase 1 offers more than 100 megawatts of critical IT capacity. Phase 2 is going to bring in another 250, and quite a bit got done this quarter; on June 3 we closed a definitive joint venture with Metrobloks. All Class A interests sit with Soluna, which also acts as manager. Regarding design, John mentioned during the Q1 presentation that Phase 1 stood at 30% schematic. At this point we are close to finishing design development. For those who aren't familiar with this terminology, schematic designs move to concept, and design development is the stage where you lock in the electrical and mechanical topology; after that comes construction documents, which are what you give a contractor to build from. We also added our general contractor during the quarter, so the team that will construct the site is present while we wrap up design; when it comes to procurement, we executed commitments with key electrical equipment suppliers for certain long-lead items. Across every AI project I've built, those items are the long poles. Placing those orders early is essential, otherwise you invite delays; on power and site infrastructure, engineering work is in progress to enlarge the Las Majadas substation by an additional 100 megawatts to support future phases, with those upgrades anticipated early next year. In addition, we executed a gas pipeline access agreement to boost resiliency. Looking at future phases, an additional 150-acre parcels are under agreement and the purchase is close to execution. Conversations have also started with county officials regarding a potential tax abatement; as for the tenant, a letter of intent has been signed, and commercial terms and lease negotiations are underway right now while we finalize the design. Once there's a definitive agreement to announce, we'll update the market; until then, the clear signal I can offer is what we're doing during negotiations. Design is finished through development. The contractor is in place. Long-lead orders are going out. Substation engineering is proceeding in parallel. That's the work a lease requires, and we're doing it now. Let's move on to our second AI project, Dorothy 3. Dorothy 3 is expected to deliver 300-plus megawatts of AI capacity on land next to our existing Dorothy campus in West Texas. As for land, 397 acres are under contract to support the initial build-out. On the design side, preliminary master planning has started and design teams are mobilized. We've also kicked off long-lead equipment procurement activities, launched environmental, water, survey and fiber studies along with advanced ERCOT integration work; on-site gas options are also being evaluated. At present we have more than 1.6 gigawatts of AI data center capacity in development. Two aspects of this quarter's expansion deserve your attention. The first is where that growth originated. 4 sites in our existing portfolio grew larger. Project Rosa went from 187 to 242 megawatts. Project Hedy went from 120 to 198. Project Ellen went from 100 to 145 and Project Fei doubled, going from 120 to 240. That amounts to an additional 300 megawatts added at sites where the work has already been done. The land is identified, the power agreements are signed, and we have exclusivity. Rosa, Fei, Hedy, and Ellen have all been set aside for AI-related work now. In addition, we assigned Project Grace—our two-megawatt technical validation initiative alongside the Siemens PTI team—to Dorothy 3 capacity. The thing we hear from nearly every investor is this: wind isn't blowing around the clock. So how can a data center sitting behind a wind farm serve a tenant requiring electricity 24 hours a day, each day, across 15 years? What solves it is supplementing those wind farms with extra generation resources that function as a single integrated system. Number one, the renewable plant itself—sustainable energy that would otherwise be curtailed and wasted, at power costs of roughly $40 per megawatt hour. Number two, the grid. Because of our behind-the-meter design, the campus can pull from the plant or pull from the grid, depending on what the moment calls for; that flexibility is exactly what brings us to power in months rather than years. Number three, firming. At Kati 2, we've signed an access agreement with an operator of natural gas pipelines, and engineering work on a lateral to the site is in progress; on-site generation that lifts the campus to the availability level an AI tenant signs a contract for. It's a truly fascinating time to be doing this work, right at the intersection where AI demand meets renewable energy that would otherwise go to waste. I'm extremely happy to be here. With that said, I'll pass things to Mike Picchi for the financials. Mike?
Appreciate that, Ryan, and glad to have you with us. My remarks will focus largely on the balance sheet and capital, so I'll begin in that area. During the quarter, our liquidity improved significantly. We closed Q2 with $113 million of cash on hand for project development and operations, and the current ratio moved up to 2.1x, compared with 1.7x at Q1. Working capital stands at a positive $69.2 million. Looking at assets, current assets climbed 48% to $134.6 million, while net property, plant and equipment increased $58.3 million to $137.8 million; that reflects the Briscoe Wind assets and Kati 1 construction being added to the balance sheet. Total assets were up 54% to $293.5 million. On the other side, total liabilities increased slightly, to $81.8 million from $76.1 million. Inside that figure, one reclassification deserves some explanation. Current liabilities rose 26% to $65.4 million, whereas long-term debt dropped 81% to $3.0 million; this is not new borrowing. What happened is that our generate debt was reclassified from long term into current, since we plan to repay it in the near term; in fact, earlier this week we prepaid slightly more than half of the outstanding loan. Total debt for the business as of June 30 came to $33.1 million. To put it briefly, when the quarter closed we had far more liquidity, a much larger base of infrastructure we own outright, and a simpler capital structure compared with how we started. Turning to the P&L: revenue reached $15.1 million, up 145% from a year earlier, which made it the fifth quarter in a row of sequential growth as sites and customers kept ramping. One accounting item touches this line. From this quarter onward, pass-through electricity costs get reported on a gross basis rather than net, affecting both revenue and cost of data hosting. This change to how we present things was only applied going forward. It adds about $4.4 million to each line in Q2, and it doesn't affect gross profit, operating loss, or net loss at all, and we haven't restated any earlier quarters. So the top line can be interpreted two different ways, and both of them are correct. Looking at it on an as-reported basis, revenue increased 145%. If you exclude the presentation change, revenue grew 73% compared to last year and 13% from the previous quarter. There's a footnote on every slide that's affected. As for what drove the growth, hosting revenue got a boost because Dorothy 2 operated for a whole quarter, unlike the year before. Kati 1 also pitched in, Dorothy 1A brought Blockware and Canaan up to speed, and Dorothy 1B began hosting. Briscoe generated $366,000 in wind revenue after the intercompany elimination was applied. On the other side of the ledger, proprietary mining declined by $1.1 million, a 40% drop, driven by a 34% fall in hash price — roughly $51 down to $34 — and by our intentional decision to move Dorothy 1B capacity away from mining and into hosting. Gross profit came in at $766,000, which is 35% lower than the $1.2 million recorded a year earlier. I want to be precise about the two causes here, especially because revenue went the other way. Expenses tied to new sites landed ahead of the full revenue those sites are meant to produce. Kati 1 is coming online in stages, so the cost of running it appears before the site reaches full operation. On top of that, Briscoe saw higher repair costs. During its initial quarter as part of our portfolio, the wind farm absorbed about $1.5 million in turbine repair and maintenance costs and recorded a gross loss of $787,000. Ryan explained the reasoning behind our decision to pull that work forward. Beneath the consolidated figure, the operating fleet did its job. Data hosting produced $1.9 million in segment gross profit, making it the biggest single contributor. Demand response came in at essentially full margin. Proprietary mining recorded a loss as hash prices compressed, while Kati 1 achieved its first positive gross profit — the threshold at which a site stops burning capital and starts producing it. I should add that cost of revenue reflects depreciation at the site level. Turning to how each segment performed on EBITDA, the second quarter of 2026 posted a 50% increase over the same period in 2025. We recorded a net loss of $22.6 million in Q2, compared with $7.8 million a year earlier. The larger shortfall reflects non-cash and financing-related effects that I'll detail shortly, with stronger results at the site level providing some offset. On an adjusted basis, EBITDA showed a loss of $1.6 million, which is 25% better than the $2.1 million loss we saw in Q1 and roughly unchanged from the year-ago figure. Stepping back further, that marks a 76% improvement relative to the $6.4 million loss posted in Q3 of 2025. What really counts here is the direction of travel. Digging deeper, the bridge from net loss to adjusted EBITDA includes three distinct components. Starting with the first item: noncash stock-based compensation came to $9.4 million. That figure reflects 2025 and 2026 equity awards that overlap, amortizing under schedules locked in when they were granted. Moving to the second item: interest expense hit $3.2 million, which is higher than the $1.2 million recorded a year before. And third: we saw a $4.2 million loss on debt extinguishment, split with $2 million connected to the Generate Tranche B modification that partially funded Briscoe, plus another $2 million connected to paying off the Yorkville loan early. On the G&A side, expenses climbed $9.8 million compared with the prior year, and $7.5 million of that total came from the rise in noncash stock compensation. Salaries and benefits added $1.1 million, while professional and legal fees added $720,000 — both driven by the Briscoe transaction, negotiations over the hosting agreement, project financing, and technical accounting support. You'll find all the full reconciliations inside the appendix as well as in the earnings release. Now I want to move on to capital — the topic I monitor the closest and the one that draws the most inquiries from people. During Q2, we brought in $159 million. Breaking that down: the ATM program accounted for $113.5 million, debt financing gave us $24.5 million, the standby equity purchase agreement contributed $18.9 million and has now been used up entirely, and warrant exercises added $2.5 million. Regarding the debt portion, we drew $12.5 million against Generate's Tranche C so the Briscoe transaction could be finished, while $12 million was provided by the Yorkville note — drawn and then completely paid back inside the same quarter, on June 12. And we put $159 million to work. A total of $51.4 million in net proceeds was allocated to Briscoe Wind Farm, while $25.3 million was used to acquire the stakes held by our joint venture partners at Dorothy 1A and 1B. Debt repayment accounted for $17 million, CapEx and equipment deposits took $13 million, and working capital plus other items came to $7.4 million. In the end, $45.3 million in cash flowed to the balance sheet. That last number matters a great deal, and there's a reason for it that I'll explain once we get to the next slide. We also simplified our capital structure. All 62,500 shares of Series B preferred stock were exchanged for 6,510,416 common shares, based on an adjusted conversion price of $0.96. Accumulated dividends totaling $2.1 million were paid out, and on June 23 we filed a certificate of withdrawal. There is no Series B preferred stock still outstanding. After the quarter closed, we issued roughly 18.8 million more shares through the ATM program, bringing in net proceeds of $23.6 million. Let me now wrap up by explaining how we plan to finance what Ryan laid out. To win an AI contract, you have to spend before that contract is in hand: fiber studies, environmental and geotechnical work, long-lead equipment, deposits, site design, joint venture formation, securing power. When a customer performs diligence on a site, they're assessing work that's already been completed. You cannot simply wait until a contract is awarded before beginning work. The way we handle a major build-out is through debt at the project level. If we're building out 100 megawatts, which we peg at between $1.2 billion and $1.3 billion, our target would be 70% to 80% in project-level debt — most likely high-yield notes, sized against the cash flows from the contracted tenant under the lease — while the leftover 20% to 30% would be funded by Soluna contributions or equity from third-party projects. Our expectation is that this debt and equity structuring would come together in the 8- to 12-week span following the announcement of a signed lease. Being able to have capital on hand to break ground during that window is yet another way we'd use the cash balances accumulated over the first 7 months of 2026. When you look at things on a project-by-project basis, we keep Bitcoin and AI in separate capitalization buckets. That way, each one can tap funding that matches the risk it carries. That's how we approach underwriting a lease for AI. Given the lease pricing we're seeing right now in the market for critical IT capacity, structured as long-duration triple net deals, a contract spanning 15 years at 100 megawatts would bring in enough revenue and net operating income to cover the project-level debt, pay it off, and still produce a robust multiple on the equity put in. The example on this slide illustrates about $180 million in stabilized yearly net operating income, set against an estimated build-out cost of $1.2 billion to $1.3 billion, using the assumptions laid out here. Once we finalize an agreement, actual outcomes could vary significantly. The reason I'm bringing this up is to show why we're putting money in before a contract is signed. This isn't a projection, and keep in mind it represents only the initial 100 megawatts of our AI pipeline, with room to grow at every campus and more projects that could be added. To wrap up, Q2 was about consolidating assets: we gained complete ownership of Dorothy 1, bought the generation facility beside it, established the Kati 2 joint venture, eliminated the Series B preferred stock, and finished with the strongest liquidity the company has ever had. Our operating fleet got better, adjusted EBITDA rose quarter over quarter, and the capital we spent became assets on our balance sheet. I'll now turn things back over to John.
Appreciate it, Ryan and Mike. I want to wrap up by describing how the quarter took shape. We expanded the pipeline in areas where we already have power in hand, growing term sheets at Rosa, Hedy, Ellen, and Fei by over 300 megawatts. Vertical integration was finished at Dorothy 1. On April 1 we picked up Briscoe Wind Farm, and after that we bought out our partners at Dorothy 1A and 1B, which let us bring vertical integration to the D1 site. Both AI campuses moved forward — Kati 2 went from design development into construction documentation, with a signed letter of intent from a prospective tenant, while Dorothy 3 progressed through land, fiber and utility studies, pipeline, power, projects and now customers. That's the order in which this business operates. And during the second quarter, we pushed ahead on each of them. Everything rests on clean energy that would otherwise be curtailed or wasted. That's the idea behind renewable computing. Our mission is to turn renewable energy into a global superpower, with computing serving as the catalyst. Thanks for spending time with us today. Operator, let's open it up for questions.
Congratulations on the impressive strides the platform has made, both in terms of execution and financing. For my first question: last month you said an LOI had been signed with a prospective tenant for Kati 2. I'm wondering whether you can provide additional detail about where those discussions stand, what still has to happen prior to a lease being signed at the campus, and what timeline investors ought to anticipate?
I'll take that one. Michael, thank you for being on the call. The lease timeline is something we hear about frequently. These leases, as you're aware, involve highly complex work, with extremely detailed negotiations required, and all of that is currently in motion. So although timing isn't something we can address, here's what we can tell you: at this point we're operating the way an infrastructure — an AI infrastructure — company does. We're assembling execution teams, developing our sites, designing the data centers, placing orders for long lead equipment, as you've been hearing, kicking off preconstruction activities, carrying out community development and preparing capital formation while we work through commercial terms with at least one prospective customer. So my response regarding timing would be to stay tuned, and I'd add that our progress in this process is solid.
One topic that keeps coming up from investors is the power required to get Soluna's AI data center projects that are close to launching up and running, Kati 2 and Dorothy 3 in particular. John, I appreciate the details you gave about those sites sitting adjacent to locations that are already energized and currently powering Bitcoin mining data centers. Still, how should investors think about the energized capacity available today that could in principle be directed to Kati 2 and Dorothy 3, and then the additional megawatts that would be needed just to run the first initial phase of each project, especially while the audit requirements in ERCOT are still being worked through? I get that this is a big ask, but it keeps coming up when we talk with investors.
Yeah. No, we — I can see why we'd get asked that. To start, like I mentioned on the call, we're fully behind the audit process. Governor Abbott has the right approach. Data centers ought to benefit the community and the grid, and they should be considering how they'll impact those two things. So, for the last eight years, everything Soluna has designed has centered on exactly that — creating a beneficial impact for the community and building data centers that work with and support the grid. Being located behind the meter works in our favor, since we're already tied in with generation assets. That's something we find exciting, because the governor seems to be suggesting that this is how it ought to be done, and it's what Soluna has been doing all along. So, to address your question, Dorothy 3 is going to draw its power from both the grid and the Briscoe Wind Farm. Like I mentioned, that asset is now ours, which means getting power over to the Dorothy 3 project is pretty straightforward. Meanwhile, the Kati platform—the Kati campus we're developing—pulls its electricity from the Las Majadas wind farm. That facility is already up and running, as we mentioned, right across the street on the Bitcoin side, which keeps us aligned with the idea that energized loads matter less to us now while study loads matter more. So we'll be drawing energy from Las Majadas, and we've coordinated with them to boost how much power we're able to take from that plant. So that's the reason we're making further upgrades—to handle that higher power requirement, so to speak, and to accommodate the campus growth. Another point I brought up during my part of the discussion, and Ryan also touched on it when updating everyone about Kati 2, is that we're putting money into delivering firming energy to the location by pulling from nearby gas lines situated close to Kati 2. That will let us explore building more generation capacity out there, which would supply extra prime power to the location, meaning both campuses end up with available energy. The two campuses already have power flowing to them, thanks to the grid integration work we funded earlier, so we believe our position is strong going into the audit. An audit is coming for everybody, and we're glad to have it. We already know the answers we'll give to those questions, and my sense is that ERCOT and the governor's office will take them favorably.
Also, congratulations on how far things have come. Earlier, you talked about interest from hyperscalers and neoclouds regarding Kati 2. Could you help us describe the potential tenant that is currently under LOI? Then, when it comes to Dorothy 3, how would you describe the caliber and the phase of the inbound interest you are receiving?
Mike, would you like to handle that one? Just spreading the love.
Appreciate that, John. Now, on the matter of getting Kati 2 interest firmed up, we haven't said if the party holding that letter of intent is a hyperscaler or a neocloud. That's all I can share on it right now. Turning to Dorothy 3, the level of inbound interest has been extremely strong; honestly, it caused us to move our development work there ahead of schedule as we move through this year. Our attention has been heavily on Kati 2, and Dorothy 3 has come up often during the past quarter. Combined, those 2 projects total 650 megawatts, yet you can already see us getting projects queued up after them, like Project Fei and Rosa; so customer demand around Dorothy 3 has sped up how we're developing that site. That has us quite excited. Having Briscoe Wind Farm in place is a major advantage. It will turn that into a truly excellent campus for us.
Yeah, and one more thing I'd mention is that the kinds of customers we're getting, plus the kinds of interest, cover the whole spectrum when it comes to hyperscale and neocloud.
Thanks for that. And Mike, you brought up accelerating work on Dorothy 3. To be clear, are you officially taking Dorothy 3 to market? Or does the plan still call for a fall start? And if fall is the start, which milestones still have to be finished prior to kicking off the formal tenant process?
I'll take that one. Our approach has been to stick with a very particular sequence we've set out, which requires completing a master plan, having started any interconnection updates and amendments required for the interconnection process for AI and land, long lead, everything Ryan brought up earlier, all before we go to market; yet with the amount of demand out there in the market at this moment, that doesn't stop people from reaching out to us. As a result, we've had inbound interest coming in, and our usual response is to share some high-level details regarding the site, and formal marketing will kick off in the fall. So I'd keep that time line in mind, though I also want to mention that our phone keeps ringing.
Got it. Ryan, congratulations on becoming part of Soluna. You brought up long lead items. Broadly speaking, how are you approaching the availability of equipment and the wider supply chain risk as you head into construction?
Regarding equipment availability, a substantial amount of critical long-lead equipment — electrical equipment — is already ordered at this point. That is consistently a risk within the industry. I mean, in every building I have ever constructed, particularly as of late, equipment is always uncertain; so staying prepared, placing equipment orders as far ahead as we possibly can as an organization, and reviewing our campus master plans while working to quantify the amount of equipment we require and the timing of when it will be needed, so that we can be ready for it; and regarding water, at present, we are focused on utilizing closed-loop systems for the most part. Direct evaporative is not being used at this time; so water at the site will not be a problem. We do have a couple of other — we do have some wells that are drilled, and we have identified aquifers capable of serving the admin and other common areas at this time.
Excellent. If I could squeeze in one more question — clearly, customer appetite and requirements will be the real driver here. I was hoping for somewhat finer detail on how data centers get designed and built out. So after a lease gets signed, what should our expectations be regarding the phasing and the sizing of the individual halls within a data center, plus the schedule stretching from lease execution to initial RFS?
Well, when it comes to how big those data halls are, that will really — it could vary based on what the client requests, correct? And at this point I'd rather not discuss anything proprietary. Speaking to that isn't something we're ready to do, of course. Could you remind me what the second portion of your question was? My apologies.
Yeah, so when we think about the schedule, when could we reach the first RFS?
At this point, for Kati 2, what we're expecting is 15 months — approximately 15 months following the contract being signed.
Regarding everything you've accomplished. I'm relatively new here. So please excuse me if this comes across as a broader question that might not fit a conference call, but I find myself amazed by the sheer number of opportunities and the level of execution you achieved this quarter, plus everything still on the horizon. What's your approach to deploying capital toward the growth prospects in front of you? And is there any consideration of possibly bringing in partners or expanding your footprint via some kind of combination, given your size relative to the market and how much you have in your pipeline. I'd really like to understand your thinking on that dynamic.
Appreciate that question. I believe we hear it quite often. We're exceeding expectations relative to our scale, so to speak; and our primary asset, our pipeline, keeps expanding, how will we turn that entire asset base into active revenue considering our scale and our ability to raise money. Our approach is the same one we've relied on from the start; we essentially take a stair-step approach to growing the company, demonstrating through each individual project that we can deliver. We'll secure funding for that first project. And then, as we deliver results and begin generating revenue, additional financing avenues open up because of our track record. We constantly evaluate both financial partners and execution partners who could assist us in scaling. That's the reason within our AI structures, we investigate joint ventures capable of speeding us up, providing expertise, funding and other resources we might lack. So when we consider the nearly 1.6 gigawatts of AI projects currently in progress, our attention is squarely on the first 2 since those will serve as the evidence that we truly are an AI infrastructure platform. We're assembling a team so we can carry out those projects. And after that's complete, we'll gain access to additional assets that support our growth and let us pursue the remainder of the pipeline. So I suppose we view this as a step-by-step process. And by taking that route, we're confident it will enable us to expand and succeed with our strategy.
That's great. My next one concerned Texas; I felt your response regarding why you're positioned well amid the present ERCOT circumstances was quite compelling, plus the case for optimism that you've essentially reached the front of the line. Based on what you've seen, does the renewable generation you provide in some way assist in speeding up your applications? And is that something that, prior to all of this occurring, you figured would be 0? Could you share any detail on that? I'd imagine that with your power profile and supplying renewables at a moment when there's substantial worry about ensuring we don't harm consumers, delivering power and renewable power is an excellent approach. So I'd really like to hear your perspective on the complexities of that ERCOT process, since you've got some background there, Ryan?
Yes. That's a wonderful question, Chris. I'll gladly address it. The way to consider it is that we've developed a tremendous level of procedural knowledge and comprehension of the ERCOT marketplace. That's among the things we excel at as an organization. And near the start of this whole large load initiative, we were among the handful of firms that met with the grid operator and supplied information and data regarding these novel kinds of data centers and technology. Here, that meant Bitcoin. And that's the insight they required in order to grasp the modeling and what would allow success through incorporating these assets, as well as how to handle the large load initiative. We're among — there were many of us, but we were among the earliest to essentially reveal our hand, so to speak. What pleases us about this audit initiative is that it extends that same commitment to openness. As an organization, we concentrate centrally on our technology and our projects genuinely serving the grid, as I noted. We set up our data centers alongside renewable energy in order to consume squandered and otherwise unutilized power, which lets the grid take in more of that kind of power. Visit any of our locations, and we're a beneficial presence in the community; we don't consume water. We generate jobs. We supply our own power, which is behind-the-meter integration. And through the AI strategy we're pursuing, we'll do additional work of that sort. And if you examine closely what occurs with power prices and such in the areas where we operate, owing to the structures and the manner in which power moves and the revenues that reach those communities, power costs in fact decline; and so — when I step back and compare the Soluna model against my peers and other loads entering the queue, we're likely a template for how this ought to be constructed in the future. So overall, we're feeling quite optimistic, since our sites already have power, we'll be part of the batch process from here on out; and we'd already been submitted as a batch load. That makes it pretty tough for a grid operator to claim you've got power, you're consuming it, and then suddenly tomorrow you don't — that isn't what this audit is trying to achieve. Its actual purpose is simply confirming these projects are legitimate and will add something to the grid. That said, we can't promise anything. This process has only just begun. I believe tomorrow is when they'll announce additional specifics about how it will function; but our expectation is that we'll get RFIs asking essentially what are you folks doing back there? In what way does this impact the community? What's your history? How will you construct these projects, time lines, budget, that sort of thing? And we're ready to supply that information, because we believe it's the proper thing to do.
Fantastic. That was a strong quarter. I'd like to perhaps pull back slightly and consider things from a somewhat different angle. Looking ahead across the coming 18 months, John, what would you identify as the one greatest danger to delivering what you've described?
I'll begin the response, then hand things over to the remainder of my team so they can share how they see it. The risk question is one I genuinely enjoy, since within the organization we pose that very thing to ourselves — what threats matter most, and what's our plan for reducing them? Our company is being reshaped into an AI infrastructure business; that wasn't what we were previously; so throughout that shift, we examine which dangers and obstacles loom largest. In my view, the top danger and obstacle comes down to how well we execute. Do we possess every tool, the architecture, the know-how, the talent, the partnerships, the capital needed to succeed across the coming 18 months? When I focus on that closely, I'd say the greatest danger earlier in this year was whether we had the right personnel, the know-how capable of reducing our execution risk and letting us operate at the very highest levels. How do you fix that? You bring in a Ryan Carver, you recruit the finest talent available and tell him to assemble a team; that's exactly what we've concentrated on throughout the latest quarter — growing our team with the know-how we now require in order to execute on this remarkable asset we hold. I'll hand the baton to Ryan Carver for his view, and I'm certain Mike Picchi has some thoughts too.
Yeah. Thank you, John. I'd say you're probably right that this is the largest exposure, though as you put it well, it's a thing we're in the process of constructing at this moment. The team is being assembled. I bring quite a deep background, extensive experience in hyperscale development at the top tier at Microsoft, and I feel very privileged to carry that experience and deliver it here to Soluna. So as John said, when it comes to risk, we're assembling a team. Somebody already brought up the risk around long-lead equipment. Getting ahead of that is how we can mitigate it, as I mentioned earlier. I'd say continuing to assemble the team — that's our largest exposure, and right now we're mitigating it.
Appreciate that, Ryan. When I look at what makes the data center business difficult, I see 2 major hurdles: getting power and getting money. The reason Soluna appealed to me so strongly comes down to that excellent power position. What I really like about the connection is this — we deliver load to power. We deliver load to generation assets that are already up and running. There's no waiting around in a line for the utility queue or for the grid operator somewhere down the road. Instead, we plug into wind farms and solar farms that are already there. That said, the second piece is the risk factor I keep thinking about — raising capital. At the project level, I believe the strategy — the plan for going after it — and securing project-level debt at roughly an 80% loan-to-cost ratio works and has been demonstrated repeatedly by — across numerous transactions during the last 5 years, it's a workable route. Then on Soluna's end, for that remaining 20%, you have to find the equity, or equity partners, or JV partners who could assist with funding the equity portion; so that's the work we do, aiming for a portfolio approach to financing these projects, which should let us carry out everything across all 1,600 megawatts currently in line of sight and active. Thanks, John.
Great. Could I ask Ryan a quick follow-up? Ryan, I'd really like to get a handle on what keeps you committed to Soluna and what has you excited about what's ahead?
To me, Soluna offers a genuinely compelling value proposition. The idea of placing a data center right next to sustainable power assets struck me as extremely, extremely compelling. And as Mike mentioned, and as John mentioned earlier, we already hold the power generation assets in certain cases and circumstances — we control the whole life cycle, all the way from producing power through to the data centers we'll be building here in the very near future. So there's a distinctive value proposition, once again, in siting data centers alongside renewable generation, which would cut the timeline dramatically. I'd be eager to build extremely, extremely quickly. Safety ranks as the #1 priority at all times, correct? That surely needs no saying when you work in construction, but we're able to move even faster since we hold our own assets, and we can bring the data center right to that spot. Plus there are multiple wind farms where we already hold pre-existing exclusivity agreements, and PPAs have already been negotiated. We enjoy plenty of different benefits, since we have lots of assets we can leverage extremely, extremely rapidly and won't be prevented from using when it comes to securing a place in an interconnection queue. So that means swift deployment, plus power consumption that's sustainable and traceable. Across the data center sector, plenty of different businesses are offsetting their environmental impact. Lots of folks are moving toward sustainable carbon zero, net zero by 2030, and similar commitments. We've got sustainably — we've got traceable, sustainable power where we can demonstrate how much we're consuming, quicker time to market, and we hold the assets. So all of those things appeal to me greatly.
That wraps up the question-and-answer portion. Today's call ends here. We appreciate your participation. You're free to disconnect at this point.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Soluna Holdings, Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Soluna Holdings, Inc. earnings transcripts and 131,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.