ERock Q3 2026 Earnings Call - Revenue Quadruples Sequentially, First Positive EBITDA, and a $1.6B Backlog Signals Structural Power Shortage

Operator: Greetings, and welcome to the ERock Third Quarter 2026 Earnings Conference Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. If anyone should require operator assistance during the conference, please press star zero on your telephone keypad. Please note that this conference is being recorded. I will now turn the conference over to Ted Durbin, SVP of Finance. Thank you, Ted. You may begin.

Ted Durbin, SVP of Finance, ERock: Thank you, operator. Hello, and thanks for joining ERock’s Third Quarter 2026 Earnings Call. I am joined by John Carrington, CEO, Ian Blakely, CFO, and Corey Amthor, President. In a minute, I will turn the call over to John and Ian for their opening comments, after which we will have a question-and-answer session. Please note that we have also published a supplemental earnings presentation on our investor relations website. Later today, a replay of this call will be available on our IR site. As a reminder, we will provide forward-looking statements during this conference call. These statements reflect our current expectations and are not guarantees of future performance and involve a number of risks and assumptions. Please review our SEC filings for a list of actual risk factors that may cause our results to differ materially from those presented during this conference call.

We will also discuss non-GAAP measures during this conference call. Reconciliations of Adjusted EBITDA and certain other non-GAAP measures to their closest GAAP counterparts can be found in our earnings release and presentation available on our website. With that, let me turn the call over to John.

John Carrington, CEO, ERock: Thank you, Ted. Good morning, and welcome to ERock’s Third Quarter 2026 Earnings Call. Our team is executing well, and the demand for our solutions is robust. There has clearly been a fair amount of volatility across AI infrastructure recently. However, our focus has been on controlling what we can control and delivering what we said we would. In 3Q, we delivered. Revenue grew roughly fourfold sequentially to $153 million, and we achieved positive EBITDA of $13 million, which represents a $27 million increase in EBITDA since last quarter. These results are in line with the outlook we have provided since our IPO, and we are affirming our full year 2026 outlook. Ian will provide more details on our results and guidance later in the call. Operationally, we made progress across several aspects of the business this quarter.

Production in our Hyperion facility ramped through the quarter, and we remain on track to exit the year at approximately 1.2 GW of annual capacity. Our El Paso Electric project moved from mobilization into active construction. We contracted for a multi-year fixed-price supply of engine long blocks from Hyundai. We secured an additional production location in Houston called Atlas, which we expect to add roughly 900 MW of annual capacity by the first half of 2028, with the ability to expand to up to 1.2 GW over time. Finally, we added over $300 million of bookings after the close of the quarter that brought our backlog to a new record. Before I walk through each of these in greater detail, I want to briefly comment on why our business model is structurally different from many companies in the AI infrastructure space. First, we’re capital light.

We don’t own assets and don’t intend to become an asset owner. As of the quarter end, we had over $500 million in cash and no debt, and we can expand our production capacity at relatively low cost. Atlas, which I will come back to later in the call, is a prime example. Second, our cash conversion is strong. We receive a deposit at contract signing and further payments as we hit operational milestones, and we match those collections with payments to our suppliers. That keeps us working capital neutral and less exposed to project delays. Third, we work with high-quality counterparties such as El Paso Electric and Meta, Anthropic, Entergy, and Microsoft. Fourth, we are building long-term recurring revenue streams with our O&M service agreements. The result is that we control our own destiny. These advantages mean that we don’t depend on capital markets to fund our growth.

Now, turning back to our results. Operationally, we made solid progress across several parts of the business since our last call. Starting with production, our Hyperion facility ramped through the quarter. In July, we moved generation production from our original Titan facility into Hyperion, expanding from 100,000 to 400,000 sq ft. After a short planned pause during the transition, output improved steadily month-over-month. We remain on track to exit 2026 at 1.2 GW per year. Moving to our El Paso project, we are partnering with El Paso Electric to support Meta’s data center complex with 366 MW of power. We are compressing what would otherwise be an estimated five-year grid interconnection wait into roughly 15 months. During the bridge period, El Paso collects payments from Meta under a power purchase agreement.

Once interconnection is complete, this highly depreciated asset can move into El Paso Electric’s rate base, giving ratepayers a meaningfully lower-cost source of peak capacity. The execution is on track. Most of our civil work is complete, and gensets began arriving on site earlier this month. We expect to complete installation by mid-2027. On the regulatory front, EPE received its air permit in under 45 days in late 2025. In September 2026, EPE also received a favorable proposal for decision from the administrative law judge regarding this project. We’re aware of the recent discussions in Texas around how large loads connect to the grid. We have seen no impact on this project. Meta has engaged directly with the community, including establishing a local fund in El Paso, and we believe the project is well-positioned and will benefit both Meta and EPE ratepayers over the long term.

While we continue to execute in El Paso, we are also strengthening our supply chain. In early October, we announced a multi-year fixed price agreement for engine long blocks, a core component of our RockBlock generator. The agreement covers 2027 through 2028 and secures supply for our production commitments already in our backlog. Locking in this long lead component early protects both our delivery schedules and our margins. It gives us a strong foundation with Hyundai as our needs grow beyond 2028. Turning to capacity expansion, our production is sold out through 2027 and into early 2028. Given that demand and our Hyundai agreement, we have secured a location for a new assembly facility in Northeast Houston, which we are calling Atlas. At 650,000 sq ft, Atlas will be 60% larger than Hyperion.

The location is close to natural gas supply for quality testing and has good proximity to the Houston Ship Channel for logistics. We expect Atlas to begin ramping production in late 2027 and reach roughly 900 MW of annual capacity by the first half of 2028, with the ability to expand to 1.2 GW over time. We expect Atlas CapEx of just $20 million-$30 million, depending on the final configuration. That’s the capital-light model in practice, roughly a gigawatt of new annual capacity for a fraction of what it would cost a traditional OEM. Together, Hyperion and Atlas will give us about 2.1 GW of annual capacity by the first half of 2028 and up to 2.4 GW with the expansion. We will apply what we have learned building Titan and Hyperion as we build Atlas, and shared overhead should lower our per-unit cost over time.

Before discussing the commercial side of the business, I want to spend a few minutes on the installation and ongoing service aspect of our model, which we believe are a competitive differentiator for the business. On our August call, we spent time on RockBlock, our proprietary natural gas generator, and its technology differentiation. RockBlock was purpose-built for fast start and dynamic transient response, with up to 99% lower emissions than diesel, quieter operation, a small footprint, and no water requirement. The generator is only one part of what we offer. We deliver a complete turnkey power system. This vertical integration allows customers to work with one accountable partner from installation through ongoing operations. On the front end, we help customers with engineering and design, technical specifications, and grid certification.

Our dedicated program managers work hand in glove with our customers to optimize time and resources to ensure they receive a quality project. We also lead negotiations with suppliers for transformers, switchgear, and other balance of plant items to minimize cost and maximize efficiency of the site. We help manage permitting, natural gas supply, and environmental studies to ensure on-time and on-budget power system delivery. El Paso is a great example. With our support, the customer received their air permit in less than 45 days. Our vertical integration drives deployment efficiency. Our pre-kitted modular systems require less specialized labor and heavy equipment on site than most competing technologies, which means faster deployment. For the California Department of Water Resources, we set 50 MW of generation in just two days and repeated that at two additional 50 MW sites.

These advantages are particularly relevant in today’s constrained labor market, where skilled labor availability can affect both project costs and schedules. Our installation model helps mitigate those constraints while supporting faster, safer, and more repeatable deployments. Once the installation and commissioning are complete, we operate and maintain the system over the life of the asset. In the field, we provide rapid outage response, preventive maintenance, and inventory management to deliver five nines of reliability. Because our equipment and processes are standardized, we do not require specialized engineers. Trained technicians can perform routine operations and maintenance, supported by centralized engineering and technical specialists. This expands our addressable technical hiring pool and allows us shared specialized expertise across the fleet. Those standards also simplify spare parts inventory needs. We also provide centralized optimization and insights, including reliability and economic dispatch, natural gas procurement, hedging, and reporting.

Our network operations center monitors every system and captures data over each genset’s life cycle, and that data helps us improve performance and drive more efficient maintenance over time. Field findings also feed back into engineering, supplier quality, and production, allowing operating experience to inform improvements to both our equipment and maintenance practices. Another good example of our approach comes from H-E-B, the largest grocery store chain in Texas, which has effectively standardized on ERock for backup and dispatchable power. We provide critical backup support during storms to help maintain power to critical operations during utility outages, and we help H-E-B monetize those same assets during grid events or when power market prices provide an attractive return. Our technology also helps customers address the environmental and community considerations associated with adding significant power capacity.

RockBlock’s low emissions, quiet operation, and lack of water requirements can support deployment closer to load centers and help differentiate our systems in the permitting process, depending on the jurisdiction. We are already seeing customers value these attributes in competitive situations, and we believe they will remain increasingly important as data center developers balance the need for power with their environmental footprint and relationships with local communities. As you know, the need for computing power continues to increase, and data centers will be built to meet that need. We believe the power industry is in a structural shortage, and our solution meets that demand for power while enabling data center operators to remain good stewards of their communities. Another component of our offering is the multiple use case for a single asset.

On our bridge-to-grid model, we are seeing bridge power periods trend somewhat longer than previously discussed, starting at three years, but increasingly pushing to five to seven years. This largely reflects the time required to secure full grid interconnection in certain markets. We view this as constructive to our model. We generate roughly three times the amount of ongoing services revenue during bridge periods versus backup and dispatchable periods. We are monitoring developments in Texas, where regulators and policymakers continue to discuss how large loads connect to the grid. We view the focus on responsible development, grid reliability, and community considerations as consistent with our approach. Whether customers are awaiting grid interconnection or evaluating other power arrangements, we offer a critical service and believe our ability to deploy reliable capacity quickly, operate it efficiently, and serve multiple use cases positions us well. Now, pivoting to some commercial updates.

On our last call, we talked a lot about data centers, so I want to spend a moment on our commercial and industrial business. Large headline agreements with hyperscalers and data center customers remain a long-term growth driver, but they can also be lumpy from quarter to quarter. Underneath that, the C&I business continues to grow steadily. One of the key drivers of the growth in C&I is the movement to reshore manufacturing. Whether it’s AI rack production and testing, drone and robotics manufacturing, or pharma production, the growth in demand from this segment often needs bridge to backup power. Legacy verticals like hospitals, universities, grocery, EV charging, and logistics continue to expand. These loads can range in size from 5 MW to 100 MW. C&I customers face many of the same challenges as utility scale and data center customers, including long interconnection queues and limited equipment availability.

We’re continuing to win share from diesel in this segment as customers increasingly move from pure backup-only use cases towards the flexible multi-use model we built our business around. A good example of the growth in demand and the value of our multiple use cases comes from Wistron, a Fortune Global 500 manufacturer. Our deployment with them in Texas began as a bridge power use case to solve the speed-to-power challenge. Once they received grid connection, our generator shifted to backup mode. Now, they are contemplating increasing the size of their facility, driving higher power needs and additional bridge power for ERock. Their familiarity with our team, execution to date, and the multiple use cases from a single asset has differentiated the company at Wistron. We also recently closed a bridge power deal with a large healthcare system for 2028 delivery.

This customer was facing a four-year interconnection wait for a major new facility they’re building. We will provide them with bridge power for a couple years, then transition into backup service. Finally, on the commercial front, in late October, Jason Altobelli joined us as chief revenue officer. Most recently, Jason served as global head of energy and infrastructure partnerships for Google Cloud, building strong relationships with Google’s key utility, IPP, and developer partners. Those are exactly the relationships we are scaling. His experience on the buyer side of hyperscaler power will be invaluable as we grow our pipeline with data center and utility customers. Taken together, Q3 was another quarter of delivering on what we said we would, and we are building the backlog, the capacity, and people to continue this momentum and execution. With that, I’ll turn it over to Ian to walk through the financials.

Ian Blakely, CFO, ERock: Thanks, John, and hello, everyone. I’ll walk through our third quarter results, our backlog, our balance sheet, and our outlook for the balance of the year. I wanted to give a quick reminder that we generate revenue in two ways. The first is power system sales, which covers the generators themselves, along with the design, construction, and installation work to get them commissioned as a fully functional power system. The second is ongoing services, which reflects the operations, maintenance, and asset management work we perform once a system is running, typically under contracts of five to 15 years. Total revenue for the third quarter was $153 million, compared with $40 million in the second quarter. This represents a nearly quadrupling of revenues sequentially, primarily reflecting ramp-up on our El Paso Meta project.

Growth at this pace reflects the execution John spoke to earlier, delivering on time and at the scale our backlog requires. Q3 power system sales revenue was $136 million, up 5x sequentially from $27 million in Q2 on the ramp-up in genset deliveries and installations for El Paso. Ongoing services revenue, excluding pass-through reimbursable variable revenue, was $8.6 million, up 23% from the second quarter, reflecting higher seasonal revenue. As we noted on our last call, some quarters will have service revenue increases due to higher levels, activity levels during certain periods of the year. The service revenue from our large projects will begin to kick in late next year and in 2028. Gross profit was $34.9 million in the quarter, representing an adjusted gross margin of 24.1%. For comparison, second quarter adjusted gross margin was 22.2%.

This overall margin expansion resulted from higher product margins, partially offset by lower installation and ongoing services margins. I want to note that we run the business on an annual and multi-year basis, but our quarterly margin will fluctuate based on which projects we are delivering and the mix between generator sales and installation work. The projects we are delivering on in 2026 and 2027 generally have margins in the high teens to low 20s. However, we have seen a structural shift to margins in the mid-20s or higher for deals that we have priced more recently. That’s a meaningful signal of our market position, but with so much near-term production already contracted, it’s largely a 2028 and beyond story. Continuing down the income statement, net income was $8.9 million, our first positive quarter as a public company.

Adjusted EBITDA was $12.6 million, compared with negative $14 million in the second quarter, representing an improvement of almost $27 million. The improvement reflects the material ramp in revenues and gross profit, partially offset by a slight increase in G&A. Contracted power system sales backlog was approximately $1.6 billion at September 30th, compared with $1.7 billion at June 30th. As John mentioned, this reflects generator production and installation activity converting into revenue on schedule, partially offset by new bookings in the quarter, including the commercial and industrial signings John referenced. We’d remind everyone that these large contracts can be lumpy quarter to quarter, and we wouldn’t read this single data point as a change in the underlying trajectory of the business. As mentioned, we signed significant contracts after quarter close, which will be reflected in backlog when we report Q4 results.

We only include firm contracts in backlog. For certain large customers, backlog reflects the generator purchase only. Installation and services agreements are typically negotiated separately, and they are added to the backlog when signed. We expect the majority of backlog to convert to revenue by the end of 2027, and our production capacity remains sold out through early 2028, even after contemplating modest volumes during the ramp-up from our new Atlas facility. We are actively working to sell out the remainder of our capacity for 2028 and are continuing to see very strong demand. Turning to the balance sheet and cash flows, as of September 30th, we had $549 million in unrestricted cash and no outstanding debt. Our $250 million credit facility remains undrawn. Year-to-date capital expenditures were $30 million.

This quarter, our capital expenditures reflected an approximate $10 million purchase of certain equipment from a customer due to a change in the use case for that site. Excluding that purchase, year-to-date CapEx was approximately $20 million. As John mentioned, because we are fundamentally an assembly operation rather than heavy manufacturing, the decision to move forward with Atlas will not require extensive capital spending. As John also mentioned, we expect capital spending in the range of $20 million-$30 million for Atlas, depending on the ultimate scope of the facility. Given our margins, we expect a strong return on this investment, and we expect to fund the Atlas expansion from existing liquidity. Our cash flow continues to benefit from upfront payments at contract signing and milestone payments as we purchase long lead time equipment and deliver finished generators and balance of plant.

On guidance, we are reaffirming our full year 2026 guidance of $435 million-$465 million in revenue and adjusted EBITDA of $3 million-$9 million. That implies fourth quarter revenue of $211 million-$240 million, which would be a new record for ERock, and adjusted EBITDA of $17 million-$23 million, reflecting continued strong execution for the business. In the fourth quarter, we expect continued growth in product and installation revenue and gross profit, largely driven by the El Paso project, partially offset by slightly lower installation margins and slightly lower seasonal service revenues. We remain focused on G&A cost discipline as well. With that, let’s turn it over to the operator to begin Q&A.

Operator: Thank you. We will now be conducting a question-and-answer session. If you would like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star two if you would like to remove them from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. We ask that you please limit yourself to one question and one follow-up question. One moment, please, while we poll for your questions.

Nick Amicucci, Analyst, Evercore: Guys, I am assuming there is some synchronality in terms of the people there.

Operator: Our first question comes from the line of Nick Amicucci with Evercore.

Nick Amicucci, Analyst, Evercore: That expense will continue. Good morning, guys. How are you?

John Carrington, CEO, ERock: Good, Nick. Good morning.

Nick Amicucci, Analyst, Evercore: I am doing all right. Hanging in. John, just wanted you to touch a little bit on the demand profile and what you guys are kind of seeing going forward. Obviously, trying to fill out the balance of the capacity that you guys are expanding on. Just any kind of conversations, whether it is, I mean, is it primarily data center customers, or are you starting to see a lot more on the utility side, just given that they are really in a bind?

John Carrington, CEO, ERock: Yeah. Thanks, Nick. I would say a few things. I think we are seeing strength across all three segments. I mentioned the prepared remarks, the C&I piece. Overall, demand very strong. We’re really not seeing any softness. I would say the larger deals are lumpy from quarter to quarter, and I know that Ian and I have discussed this with many of you. I think the way to think about a single quarter’s bookings really won’t capture the trend. I would say that since the midterms, customers have actually come back to the table with additional interest. We think that’s in large part because post-midterms, there’s more clarity, and I think there’s a lot of constrained projects that still need power but have the permitting. So, we’re monitoring that closely.

As I’ve discussed with many of you over the last couple months, the other thing to highlight is that we really expected customers trying to contract in the second half of 2026 would find little capacity remaining for the 2027 deliveries, and that’s exactly what’s going on. In fact, we are predicting now that, and we’ve said this before, that the 2027 constraint will further exacerbate 2028, and I think there’s going to be another significant supply-demand imbalance in 2028. As a result, I would say the pipeline’s now focused squarely on 2028, Nick, and increasingly in 2029 deliveries. We’ve also passed on a lot of 2026 and 2027 simply because we didn’t have the capacity, which I think really tells you where the demand is relative to supply.

Nick Amicucci, Analyst, Evercore: Great. That makes perfect sense. Thanks, John. Then, actually, just kind of piggybacking on that, if I could, because I know you guys expanded the supply agreement with Hyundai for the engine blocks. How far I mean, I guess, at what point, how strong do you feel on that supply agreement that they’ll be able and willing to scale alongside the capacity growth and just given the strong demand profile?

John Carrington, CEO, ERock: Yeah. A couple things on that. I’d say the Hyundai relationship is very strong. It’s one of our longstanding partnerships in the business. We’ve worked together for over 10 years. We recently hosted their leadership team at Hyperion, and I think they came away very impressed with our model and how we’re executing. On pricing and the contract, the contract is fixed price. I think it gives us great clarity and visibility into generator costs through 2028. Again, this was a 2027 and 2028 contract. I think it also supports our margins and really lets us have more clarity around quoting customers with confidence. Beyond that, we’re talking with Hyundai about 2029 and later. Those talks are driven primarily by the demand that we’re seeing and, quite frankly, growing interest in the larger, longer-term framework agreements from some of the data center hyperscaler customers.

Nick Amicucci, Analyst, Evercore: Perfect. Thanks so much, guys.

John Carrington, CEO, ERock: Thank you.

Operator: Thank you. Our next questions come from the line of Mark Strouse with J.P. Morgan. Please proceed with your questions.

Bhairav Manawat, Analyst, J.P. Morgan: Good morning, team. This is Bhairav Manawat on behalf of Mark Strouse. Thanks for taking our questions. My first question is regarding customer activity following the midterm elections. Have you seen any change in pipeline conversations or deployment timing for the existing backlog? I think you mentioned El Paso remains on track, but are you seeing any shifts across the other projects?

John Carrington, CEO, ERock: Not really. I can let Corey jump in, but I think if you think about the midterms, I would bucket it really into three themes that we are seeing. First, there is a big community fit, and data centers have to be built in a way that will really support the communities they operate in. I think ERock checks the boxes communities care about. We do not use water. We have a very small footprint, quiet operation, low profile, and low local emissions, all of which seem to be very much in focus amongst the new governors. The new governors also want growth and economic development, and they want communities treated fairly. I think we squarely align around that initiative. On the second one, some proof points related to that.

El Paso is probably the clearest example, and as I noted in the prepared remarks, we really are turning power systems into grid assets that will benefit ratepayers. We think this story will unfold very nicely. EPE and Meta, that project is on schedule, and we believe it is a very strong validation. Finally, our turnkey offering will matter more and more as we go forward. We really help our customers navigate uncertainty from design through operation. So, we feel good about the El Paso projects, and Corey can talk a little bit more about that. I would also add just around the momentum, which you touched on.

I would say since the election, we have seen inbound interest picked up, as I mentioned to Nick, both for new deals and for customers looking to really speed up existing conversations, and we expect to see more of that over the next few weeks. I think the market will certainly select higher quality projects, and those are the ones that will get built. I think the proof point we have in El Paso is going to resonate with every developer and hyperscaler out there to say, "Okay, here is someone that is built a prime project that will be in prime for at least four years." By the way, it is five years ahead of schedule, and we are going to put this in the ground in 15 months.

Finally, I just do not think that data centers. There is a lot of data centers with permits but no power solution, and that will give us even more momentum as we see that coming forward, and we already are. So, anything you would add, Corey?

Corey Amthor, President, ERock: I think you covered a lot. The only thing I would say is that I think it’s becoming pretty clear after the election that both parties are pro-business. As long as you’re covering those things for the local community and meeting those things that John talked about, that’s going to be very important for moving forward on these projects. We’re seeing states coming out now saying that they would take any extra data centers that they could find if other states aren’t willing to do it. That’s going to create competition also. But it really comes down to execution. I think we’re proving that we can execute, and I think our customers are hearing that and seeing that.

Bhairav Manawat, Analyst, J.P. Morgan: Awesome. Super helpful. I guess my second question is regarding the $300 million worth of bookings that you added after quarter end. Could you discuss the customer mix there and the timing of those orders as well as if those orders came from existing or new customers? Thank you.

Ian Blakely, CFO, ERock: Yeah, thanks for the question. The short answer is that it is a mix. We saw some sizable C&I deals for the trends that John was noting in his prepared remarks, and so that’s a component of it. There is also a component of it which reflects, actually to your earlier question, customers making financial commitments to preserve schedule on projects where we are still finalizing certain of the contractual details. I think that actually helps give color that customers are actually looking to keep projects fully on schedule. They’re not looking to delay or push out projects. I think that’s a significant component as well. I would say it’s a good mix of both on the data center side and on the C&I side of the business for that $300 plus million.

Bhairav Manawat, Analyst, J.P. Morgan: Super helpful. Thank you. I’ll turn it back.

Operator: Thank you. Our next question comes from the line of Joseph Osha with Guggenheim Partners. Please proceed with your questions.

Mike Stratoti, Analyst, Guggenheim Partners: Hi, thanks for taking the question. This is Mike Stratoti on for Joe. Just kind of diving into margin drivers for 2027. Ian, last quarter you said that higher prices on new orders mostly show up in margins from 2028 since you are sold out through 2027. Just kind of curious, what drives the 2027 margin gains that we should be looking at for next year?

Ian Blakely, CFO, ERock: Yeah, 2027 margin is largely driven by a mix of what we are doing at what given point in time. As noted, everything we are delivering in 2027 was priced in 2025. Then I would say a little bit of economies of scale on the assembly side of the equation. But as you noted, really, the big drivers are for 2028 and beyond. I think we talked a little bit in the prepared remarks about margins that we are seeing lower in the range in the low 20s. But as we look to 2028 and beyond, we are in the mid-20s or higher in the deals that we are currently pricing.

Mike Stratoti, Analyst, Guggenheim Partners: Great. Thank you.

Operator: Thank you. Our next question comes from the line of Angel Castillo with Morgan Stanley. Please proceed with your questions.

Angel Castillo, Analyst, Morgan Stanley: Hi, good morning, and thanks for taking my question. Maybe just to kind of follow up on that, I guess, wanted to the extent that you have a little bit more color on 2027. I think you mentioned now your backlog covers through 2027 into a little bit of 2028, but also even with your new capacity additions, I think it still only has that kind of coverage. Are you seeing your 2027 EBITDA versus prior expectations moving higher? Just any preliminary thoughts as to how we should think about the kind of 2027 just total EBITDA and how that’s kind of progressing in terms of expectations?

Ian Blakely, CFO, ERock: Yeah, so I think in general, what we’ve said is 2027 is largely locked in, so there’s not that much leverage associated with 2027 to take that beyond expectations. Really, the leverage points are in 2028. There, I think we’re seeing upside on pricing and on the cost side of the equation based on the contracts that we’re signing and negotiating, and also on what we’ve been able to do on the cost side and keeping costs well under control.

Angel Castillo, Analyst, Morgan Stanley: Got it. Thanks. One minute.

Ian Blakely, CFO, ERock: Go ahead.

Angel Castillo, Analyst, Morgan Stanley: Go ahead, sir.

John Carrington, CEO, ERock: I was just going to say, I would add a couple comments around Atlas, specifically around demand and opportunity side. I think on the demand front, seeing a strong pull across our customer base, and we really wanted to put capacity and plan ahead of it rather than chasing it. On the opportunity front, Atlas really gave us a rare chance to look at a site with access to strong gas and good logistics, as I mentioned in the prepared remarks, which is hard to find and I think very valuable for both assembly and our testing and R&D efforts going forward. Another thing I’d highlight is we’re also considering additional tooling and automation, which would, I think, raise the upfront cost but improve efficiency on margins over time.

Either way, it’s low CapEx for the capacity it adds, and we’re funding it entirely from equity, which is always a positive. We know, again, as we’ve said many times, there’s no need to return to the capital market. We’re very excited about what Atlas can bring.

Angel Castillo, Analyst, Morgan Stanley: You actually read my mind. That was actually what I was going to ask you next, is just the strategic thinking. I think in the past we had talked about Hyperion 2 and perhaps more of a 1.2 GW. I was just wondering whether there is anything there to read into in terms of the strategic choice of moving forward with 900 MW versus 1.2, or also just on that CapEx front, you touched on it. As I mentioned, you read my mind here, but the $30 million in CapEx, I guess, is higher than the $15 million per gigawatt that I think I recall in the past. Is it just that incremental tooling that you talked about and efficiencies?

Is it more, is this the level we should think about, or are there any changes here in terms of inflation as well, or just implications to the type of CapEx per gigawatt that you would anticipate going forward?

Ian Blakely, CFO, ERock: Yeah, no, I appreciate the follow-up. What I would say is on the 900 MW versus 1.2 GW, I think what we have said all along is we are going to be very prudent in how we look at capacity expansions. I think what we have done is moved ourselves forward in a way that allows us to effectively have the ability to do 1.2 or potentially even more. What we are saying is today we are committing to 900, and we will commit to the incremental as we get the incremental contracts fully signed up. I think we feel like we have preserved the flexibility, and that is all relative to, I think, what we had shared pre-IPO of a 600 MW expansion capacity. I think, again, it is a continued sign of growth and demand. From a CapEx perspective, I would say there is a little bit of general cost inflation.

There obviously are some things you do every time you build a facility that you realize investments you could make that can increase efficiency. Then I think probably the biggest swing is what John highlighted around, do we do a little bit more automation and tooling that allows you to lower cost over time? Again, we are still finalizing design, et cetera. That is, obviously, we are only going to do that to the extent there is a very strong payback period associated with it.

Angel Castillo, Analyst, Morgan Stanley: Very helpful. Thank you.

Operator: Thank you. Our next question comes from the line of Michael Sullivan with Wolfe Research. Please proceed with your questions.

Michael Sullivan, Analyst, Wolfe Research: Hey, good morning.

John Carrington, CEO, ERock: Morning.

Michael Sullivan, Analyst, Wolfe Research: Hey, John. Hey, John, wanted to ask a little more if you could expand on your commentary around the bridge power periods extending out towards five to seven years, and anything you’ve done to lock that in thus far, those longer periods?

John Carrington, CEO, ERock: Yeah. Thanks, Michael. A couple things on that front. We contract with our customers in a way that regardless of how long they’re in bridge, they can continue to extend that. From a modeling standpoint, we’re 2x from a revenue standpoint, we’re 3x in bridge versus outside of bridge. So it’s accretive to our model, and I think we’ll probably see more and more of that. Again, we talked in terms of three years, now we’re seeing more like five to six, seven even. But we’re indifferent as far as from a contracting standpoint.

Michael Sullivan, Analyst, Wolfe Research: Okay, great. Then another thing you mentioned I was hoping you could expand on was just maybe permitting issues that others may be having, and where you can be opportunistic. Any specifics there you want to expand on?

Corey Amthor, President, ERock: I can answer that. One of the things that our low emissions has really made a big difference in our low water usage at sites. At El Paso, we were able to get an air permit in less than 45 days, and we’re looking at a lot of sites that are approximate size of that or even a little bit larger where we can get the same standard air permit. So I think that’s a big advantage that a lot of the people can’t do. So customers are definitely excited about that.

Michael Sullivan, Analyst, Wolfe Research: Okay, great. Thanks so much.

John Carrington, CEO, ERock: Thank you, Michael.

Operator: Thank you. As a reminder, if you would like to ask a question, please press star one on your telephone keypad. Our next question comes from the line of Ross Fowler with Bank of America. Please proceed with your questions.

Ross Fowler, Analyst, Bank of America: Morning, John. Morning, Ian. How are you?

John Carrington, CEO, ERock: Morning. Good. Good morning, Ross.

Ross Fowler, Analyst, Bank of America: Just digging in a little bit to the backlog question and sort of following on to the earlier question about the midterms. Can you just kind of update us on a couple of pieces of that? I think you’re expecting a CCN for the Meta site here in December. The first question is related to that, and what happens if that slips into sort of Holland’s administration where she kind of wants an EO on a moratorium? Where does that sit? Then the second one is sort of on Anthropic. Where’s the air permits for that project that was announced in August? Because the TCEQ, I believe, is still paused, although we may get that opened back up in this interregnum period before Hinojosa takes the governor’s mansion in July, or moves in in July.

Just kind of where we are in the permitting process of that and kind of what the risks around that are, given sort of the pauses that we’re seeing or moratoriums or whatever you want to call them, post the midterms.

Corey Amthor, President, ERock: Sure. On the CCN, right now, we don’t see any delay in that. We feel like that’s moving forward from everything that we’re seeing and don’t see really much risk in that at all. If it did move into the next governor’s reign, I guess it would be considered the same thing that we’ve been talking about is the low emissions. Meta is working very much with the public. They put a fund together to support local communities. Our emissions, low water usage, our ability to help lower rates after the project gets connected to the grid are all supportive of that. On the Anthropic deal, we haven’t named a site on that yet. We’re still working through that, but we feel like we don’t see any issues on the moratorium.

We think there’s a window of getting projects approved, and even after the fact, under the new governor, that she’s pro-business, everything we’re hearing, and that as long as you’re doing these things that you should do for the local community, which all of our customers are doing, and also we believe is important too, will help get projects approved on time.

Ross Fowler, Analyst, Bank of America: Okay. Thank you.

John Carrington, CEO, ERock: Thank you, Ross.

Operator: Thank you. Our final questions will come from the line of Tim Moore with Clear Street. Please proceed with your questions.

Tim Moore, Analyst, Clear Street: Thanks a lot. Appreciate it. It was really nice to hear about the added capacity in Houston and Atlas, and it makes sense to really do that and have more, because you are going to have demand through 2028 coming. I just wanted to ask about, when you get a big order, let us just maybe say Anthropic, how long does it usually take for them to decide to maybe add the other revenue stream, if they do, the design, installation, and service? Because that might be, I do not know, $200 million more to your backlog. I am just kind of curious, if the site has not been named yet, does it come shortly after that? If you can just add color around something like that would be great.

Ian Blakely, CFO, ERock: Yeah, absolutely. In general, what I would say is, on all of the deals that we are working on, the expectation is that we will provide the full set of services, including balance of plant installation, selling them a fully functional power system, as well as the ongoing services once the systems are commissioned. That is the nature of the conversations that we are having. For a number of reasons, what we call the ESI contract, which governs balance of plant and installation, can take between, I would say, three to nine months after the equipment purchase agreement is signed. That largely is because, one, obviously, you need to have the final site selected before you can contract around that side of the work.

The other piece of the equation is you need to have the final design for the site selected, because we are doing that on a fixed-price basis, given the scope. You really need to go through that process of design iteration, which we have found with our customers can take a lot of iterations. You can be 15 or 20 iterations into design before you finally get it locked down. All of that becomes exhibits to the ESI agreement and does drive a little bit longer timeline associated with it.

Tim Moore, Analyst, Clear Street: No, that is terrific color. I will go with the three to nine months. It completely makes reasonable sense. The other question I had, it was nice, the pre-commentary on the Wistron customer coming back to you. If you can kind of speak about this, maybe not just them, because you already did. You start out the bridge power, they get turned on by the grid maybe three or four years later. Do you have a growing visibility on some of these other customers that you have proven they are reliable, but now they are building out whatever it is, more needs for electricity, and they need more bridge power, and they are going to switch from backup power, and three times the revenue to use bridge power?

Do you have kind of a wave of that, or a sales team that keeps in touch with those folks to see if they are ready to add more, kind of go back to some more bridge?

John Carrington, CEO, ERock: Yeah, we do, Tim. We have a specific C&I sales team, a leader for that group. I would say, yes, we do stay very close to those existing customers. As you know, our NOC monitors every site that we have in place, so there’s a constant connectivity with our customers. Yes, we are seeing more expansion. The reshoring that I mentioned in the prepared remarks is a real thing. As they expand, they’re in the same quandary they were when they first launched their facility. This is not Wistron’s an example. There are many others that we’re working with right now. The other thing is, when they think they’re going to get grid interconnection, we’ve seen delays. Someone will say two years. The two-year mark comes up, and actually there’s another two or uncertainty.

They’ll need to put additional capacity from ERock in for that alone, just to manage their growth that they had expected to have grid interconnection in the two-year timeframe that was previously committed. It’s interesting. It’s similar dynamics that we’re seeing obviously on the data center side, but very acute at C&I as well. I would note that it’s, as I said, five to upwards of 100 MW. These are big, big sites, particularly if you’re in that drone, AI infrastructure space, and some of the other large manufacturing components that, like a Wistron is a good example of in the AI space. It’s a theme that we’re seeing more and more, and we’re excited about it, quite frankly, and very well-resourced around that segment.

Tim Moore, Analyst, Clear Street: No, that’s a terrific catalyst. Thanks for sharing and elaborating. That’s it for my questions.

John Carrington, CEO, ERock: Thank you, Tim.

Operator: Thank you. This does now conclude our question and answer session. With that, I would like to turn the floor back over to John Carrington for closing comments.

John Carrington, CEO, ERock: Great. Thank you again for joining. In summary, I would say, look, we delivered. Revenue and Adjusted EBITDA came in line with the outlook we have given since our IPO. Importantly, we are affirming our full-year guidance. We are building the foundation to keep delivering. That means securing the long block supply with Hyundai, and Hyperion continues to be on track to deliver the 1.2 GW that we have discussed previously. Atlas will take us to over 2 GW of annual capacity in 2028. Demand is extending and not slowing. Customers are planning into 2028 and beyond. Bridge periods that we just talked about and throughout the call are lengthening, and the C&I is growing alongside the data center segment. Execution is very much on track.

We are making good progress executing on the El Paso project, as discussed this morning, to deliver power many years ahead of grid connection and continue to advance our projects in our backlog. Finally, we really continue to control our destiny. We have the balance sheet and cash flow to fund our growth without relying on the capital markets. Before concluding, as always, I want to recognize our team. The execution across every part of this business this quarter was excellent, and I want to thank everyone for the work that has gone into making this quarter what it was. I am excited about all the progress I see as we close out the year and move into 2027. Looking forward to speaking with all of you on our next call to discuss the 2026 full year and the 2027 outlook.

Operator: Ladies and gentlemen, thank you for your participation. This does conclude today’s teleconference. Please disconnect your lines at this time, and enjoy the rest of your day.