管理層發言
Good morning, and welcome to Fluence Energy's Third Quarter Earnings Conference Call. Joining me on this morning's call are Julian Nebreda, our President and Chief Executive Officer; and Ahmed Pasha, our Chief Financial Officer. A copy of our earnings presentation, press release and supplementary metric sheet covering financial results, along with supporting statements and schedules, including reconciliations and disclosures regarding non-GAAP financial measures, are posted on the Investor Relations section of our website at fluenceenergy.com. During the course of this call, Fluence's management may make certain forward-looking statements regarding various matters relating to our business, including, but not limited to, statements related to our future financial and operational performance, future market growth and related opportunities, anticipated growth and business strategy, liquidity and access to capital, expectations relating to pipeline, order intake and contracted backlog, future results of operations and impact of the One Big Beautiful Bill Act, projected costs, beliefs, assumptions, prospects, plans and objectives of management and the timing of any of the foregoing.
Such statements are based upon current expectations and certain assumptions and are therefore subject to certain risks, uncertainties and other important factors, which could cause actual results to differ materially. Please refer to our SEC filings for more information regarding these risks, uncertainties and important factors. You are cautioned not to place undue reliance on these forward-looking statements, which speak only as of today. Also, please note that the company undertakes no duty to update or revise forward-looking statements for new information. This call will also reference non-GAAP measures that we view as important in assessing the performance of our business, including adjusted EBITDA, adjusted gross profit and adjusted gross profit margin. A reconciliation of these non-GAAP measures to the most comparable GAAP measure is available in our earnings materials on the Investor Relations website. Following our prepared remarks, we will conduct a question-and-answer session with our team. Thank you very much. I'll now turn the call over to Julian.
Thank you, Chris, and welcome to everyone joining us today. Turning to Slide 4. Today, I will provide an update on the progress we have made in driving new order intake and building our backlog, both of which were at record levels this quarter. I'll discuss our growing business, which includes robust demand from our core customers, combined with a rapid expansion of data center customers, from which we received our first orders and contract awards totaling $850 million. We believe that the momentum of the past few months will continue in the quarters to come, driven by our differentiated product offering and our team's long-standing ability to meet customer needs. Following my remarks, Ahmed will review our financial results for the quarter and our outlook for the remainder of the year. Starting with key highlights for the third fiscal quarter. First, we signed $1.44 billion of orders during the quarter, which is nearly triple the $509 million we signed in the same period last year.
Second, included in our record order intake was our first deal with a data center developer worth $300 million. During July, we were awarded an additional $550 million of business across multiple data center sites by one of the hyperscalers that we discussed last quarter. Third, we ended the quarter with a record backlog of $6.4 billion, representing 14% growth over the second quarter and more than 30% growth since the third quarter of last year. Fourth, we ended the quarter with total liquidity of approximately $860 million, in line with our expectations. Fifth, Ahmed will discuss our third quarter financial results shortly, but revenues were affected by delays in expected project deliveries driven by the ramp-up of two new contracted manufacturing facilities. Accordingly, we are lowering our guidance midpoint for 2026 revenue and adjusted EBITDA to $3 billion and negative $10 million, respectively.
We do not take this reduction lightly and have instituted changes in an effort to ensure we deliver on our growing market demand. I will detail our plan further in a moment. Please turn to Slide 5 for more detail on our order intake. With $2.7 billion now signed to the third quarter of this year, our orders are 80% higher than the amount from last year, with utilities and IPPs making up approximately 90% of this total. We expect fourth quarter orders will be another record level for the company, and we see reason for this strong momentum to continue in future quarters given our current demand and competitive position. Please turn to Slide 6 as I detail our progress with data center customers. Our announcements on last quarter's call that we have signed two master supply agreements with hyperscalers raised our profile with other potential data center customers. Overall, our data center pipeline has increased to 16 gigawatt hours, representing a more than 35% increase compared to the second quarter.
Our pipeline now includes a mix of projects from both hyperscalers and data center developers. During the quarter, we signed a $300 million order for a behind-the-meter project with a developer. We were introduced to these customers by one of the hyperscalers we have been working with. The sales cycle for this customer was much faster than our traditional market segment, converting from lead to order in three months. We continue to see the developer segment center on speed to power solutions, and we are pleased to be positioned to meet their needs. Hyperscaler customers continue to focus on quality of power solutions, where we also stand out in terms of our ability to deliver. We were pleased to receive approximately $550 million of awards under one of our MSAs in July. These are not yet purchase orders, and we expect this will add to our total of signed orders in the coming months. These data center customers have a pipeline of projects that we continue to believe we are well positioned to bid on, and we look forward to expanding our business with them in the near future.
Please turn to Slide 7 as I discuss backlog and pipeline growth. Our backlog has benefited from record orders in two of the past four quarters and sets a strong base of revenue growth in fiscal '27. As of June 30, approximately $2.2 billion of our $6.4 billion backlog is expected to convert to revenues in fiscal '27. This compares to the $1.5 billion of fiscal year '26 revenue coverage we had as of June 30, 2025. Turning to our pipeline. We exited the quarter at $33.1 billion, which is an increase of $1.6 billion compared to last quarter. This indicates $3 billion of new opportunities after considering our conversion of pipeline into orders during the quarter. We continue to see a growing percentage of our pipeline coming from the U.S. market compared to previous years, mostly attributed to the data center segment. Please turn to Slide 8 for details on the expansion of our supply chain. We have been expanding our supply chain capacity to meet the strong demand for our products as reflected in the growth of our backlog.
New larger contracted manufacturing facilities globally are expected to increase our capacity and also deliver the quality our customers expect. A major driver of our revised revenue expectation for this year is attributable to ramping up production at two of these new factories. In the U.S., we will be the off-taker of a new fully automated facility located in Houston with expected capacity of 15 gigawatt hours per year. Completion of this new facility has been delayed by a few months due to delays in construction and issues related to the automation equipment. Limited production commenced this quarter, and our manufacturer is taking steps to address outstanding issues. We expect the facility to reach full production levels during our fiscal first quarter of '27. I will highlight this contract manufacturer has been our main enclosure supplier from Vietnam, which is a very similar facility to this new one in Houston.
We believe their knowledge and experience will be helpful as this factory moves towards full production. Our new international facilities are now fully ramped and our product is being shipped to customers on a delayed timeline because initial production did not meet our quality expectations. Corrections were implemented, and we have resumed shipping high-quality products to projects all around the world. Given the importance of timely, consistent and high-quality production to our business, we recently made organizational changes to ensure more direct oversight of and accountability for our production capability. Today, we announced that Roman Loosen will assume leadership of our supply chain; and Peter Williams will concentrate on product, with both leaders reporting directly to me. Roman currently serves as our Chief Enterprise Operations Officer and brings more than 20 years of global leadership experience at Siemens, where he held senior operational and business leadership roles with responsibility for supply chain, manufacturing and business transformation.
Roman will lead a set of managers with deep experience and skill sets in supply chain and manufacturing that have joined our company over the past few months. I am confident that this new management team will strengthen our supply chain and manufacturing to meet the growing demand for our products. When combined with our supply of domestic sales, we expect the Houston facility will expand our annual capacity for domestic content significantly compared to our current footprint. Once it is fully run and added to our current supply chain, we expect to have capacity to meet our current backlog of projects and confidence to meet the growth of the U.S. market. Please turn to Slide 9 for details on how we are differentiating in the current market. We have been successful in growing our backlog and penetrating the new and important data center customer segment in a very short period of time. Fluence has new and repeat customers who appreciate our advanced product designs, leading energy density and focus on total cost of ownership.
In addition to these factors, our proprietary software stack, including our operating system is designed to enable our customers to optimize their solution over its long-term life and allow for remote monitoring. These features can increase availability and extend the life of our solution for customers in all use cases. Specifically for data center customers, the ability of our operating system to efficiently help smooth loads and handle periods of low voltage have contributed to new awards and orders. Smartstack has been gaining favor in terms of orders this year, representing 75% of our orders year-to-date. One of the attractive features of Smartstack is that we design it as a product platform with the ability to upgrade over time. During the quarter, we announced the first evolution with Smartstack 10, which increases density of each unit from 7.5 megawatt hours to 10 megawatt hours. The ability to upgrade our Smartstack offering over time with speed and efficiency allows us to quickly adapt to evolving customer needs, which is valuable for both Fluence and our customers.
To conclude, we believe we have the right product and team to win in this rapidly growing market. With our first data center awards adding to our record backlog and a growing global supply chain size to meet future growth, we are committed to delivering for customers and creating long-term value for shareholders. With that, I'll turn the call over to Ahmed to discuss our financial results and outlook for the rest of this year.
Good morning, everyone. While our results this quarter were disappointing, the challenges we experienced were primarily related to construction and production start-up delays at new manufacturing capacity and scaling of our new products. We have taken actions to address issues and improve execution and are now tracking to our revised production plan. Importantly, these investments strengthen our supply chain globally and position us to support our growing backlog. As these new facilities move beyond the initial ramp-up phase, we believe Fluence will be better positioned to deliver profitable growth and create shareholder value. Starting with Slide 11. We generated Q3 2026 revenue of $650 million, up 8% year-over-year. This was approximately $90 million below the expectations we discussed on our last quarterly call. This shortfall was primarily driven by production delays at two new contract manufacturing facilities that are currently ramping.
Production from the enclosure manufacturing facility in Houston was pushed by a quarter due to construction and automation delays. The facility has begun limited production and is expected to achieve full production levels in the first quarter of fiscal 2027. The other issue occurred at one of our two new facilities in China, where initial production of components of Smartstack did not meet stringent standards and required rework. The facility is now producing consistently to our standards and has achieved full production in the fourth quarter. While we expect to realize the revenues associated with the Q3 projects that were delayed, the slower ramp-up compresses the timeline for production in the fourth quarter, pushing a portion of previously planned 2026 deliveries into fiscal 2027. Our Q3 adjusted gross profit reflects the lost margin from revenue shortfall and approximately $15 million of costs associated with new product rollout and production delays.
In addition, we recorded a $15 million loss on a planned battery supply agreement, most of which was associated with a single project. Despite the upfront cost, this arrangement secures the long-term supply and attractive pricing, strengthening our ability to support growing demand and price future orders with greater confidence. Turning to Slide 12 for our fiscal 2026 guidance. We have revised our outlook to reflect our updated expectation for production through the end of this fiscal year. More specifically, we expect revenue in the range of $2.9 billion to $3.1 billion with a midpoint of $3 billion. The approximately $400 million reduction versus the prior midpoint is largely the result of manufacturing ramp-up delays that pushed revenue recognition into 2027. In terms of EBITDA, we now expect adjusted EBITDA of negative $30 million to positive $10 million with midpoint of negative $10 million compared to our prior midpoint guidance of $50 million.
While there are several puts and takes relative to our prior guidance, the $60 million reduction is largely explained by two items: about $44 million of lost margin from shift of approximately $400 million of revenue into 2027 and $15 million related to the proposed long-term battery supply agreement discussed earlier. And we are maintaining our expectation for annual recurring revenue of approximately $180 million by the end of fiscal 2026. Turning to Slide 13 for an update on our liquidity position. We ended the third quarter with total liquidity of approximately $863 million, which includes approximately $365 million in total cash. Consistent with what we said on the last call, we expect total liquidity will return to the $900 million level by fiscal year-end, driven by execution on our backlog of projects included in the guidance. Bottom line, our liquidity position continues to support our near-term working capital needs, particularly heading into our highest revenue quarter.
Regarding liquidity needs for 2027, we are comfortable that our existing liquidity has us well positioned for success. That said, as Julian noted, our expected order intake is reaching new highs and to support that growth may require an additional $300 million to $500 million of working capital over the coming year. We will remain disciplined and pursue financing only where there is a clear line of sight to profitable growth and shareholder value creation. In summary, while we have more work to do, demand remains strong, our backlog continues to grow and we are taking actions needed to improve our execution and support long-term profitable growth. With that, I will turn the call back to Julian.
Thanks, Ahmed. Let me close with a few key takeaways. First, order momentum continues. Our third quarter order intake, our record $6.4 billion backlog and the initial order with the data center developers all evidence our successful product strategy and sales effort. We currently expect orders to reach a record level again during the fourth quarter of this year. Second, additional production capacity. We're adding new contract manufacturing capacity in the U.S. and abroad and have realigned our organization with new leadership to strengthen execution. Third, product offering. The integration of Smartstack density, safety and reliability metrics with our software and controls capabilities allowing for fast response, load smoothing and remote operation puts us in a dominant position to meet the growing demand of the diverse customer segments we serve. In conclusion, we are positioning our company to continue profitable growth and to deliver value to our customers and shareholders.
At this time, we will open the call for questions. Your first question comes from the line of George Gianarikas from Canaccord Genuity.
George, before you ask, I really want to apologize for the technical mishap we had this morning. We'll figure out what it is; you were waiting and we were not connected. I'm sorry for that, everybody. We really appreciate and value your time, and we know it was a little bit of a waste of time. Great. George?
分析師問答
Maybe first, if you could provide some additional granularity around the production delays and just sort of go into a little bit of detail what's happening in the facility?
Yes. Great. As we scale up the company, we are increasing our production capacity. As part of that, we brought in two new manufacturer groups, one to serve the international market and one to serve the U.S. market. On the international side, we're working with reputable, well-known and seasoned contract manufacturers. One of these manufacturers is producing our Smartstack for the international markets. One of those manufacturers producing the parts that go on top of the skids had initial production that did not meet our stringent testing, and we had to significantly delay production to ensure we got production in line with our quality standards. That caused significant delays during the quarter. Since then, we have fully resolved the issues and now they are producing full time. They are fully ramped up and doing a great job, but we won't be able to recover all of the volumes we lost during the quarter and the year.
We feel confident today that they can meet our quality and volumes going forward, which will put us in a good position to serve the international market with a competitive product. The U.S. situation is slightly different. In the U.S., we are partnering with our contract manufacturer that operates out of Vietnam to establish a fully automated facility, an improved version of the one we have in Vietnam. It is a much more automated facility to account for U.S. labor cost structures. They experienced construction delays, including utility connection delays. We have been running parts of the plant on generators, which prevents certain parallel works and slowed automation integration. Those delays compounded, and at one point it became clear we would not be able to recover the volumes for fiscal '26 and that some volumes would shift into fiscal '27. That facility is ramping up and producing today.
It will connect to the grid in the next couple of weeks. The issues are being resolved. As I said, this is a sister company to the one in Vietnam, so we are confident the production levels we have set for the quarter will be met and that the issues we identified are fully resourced and resolved. This facility will provide us a competitive advantage in the U.S. market. It will allow us to produce 15 gigawatt hours of fully U.S.-made products with fully automated integration. We are very pleased with what we will receive, however we could not fully resolve the timing to avoid the delay into fiscal '27.
And maybe just as a follow-up, an update on your recent commercial traction in data center, specifically, how would you characterize the competitive dynamics and your win rates and deals you participated in? And what are the key differentiating factors that lead to your wins?
Very good question. How we win: we win because of the density, safety and reliability of Smartstack combined with our operating system, which allows for very efficient load management and strong response times for low-voltage ride-through events. It's a combination of the technical capabilities of our operating system and the reliability and safety of our enclosures. That's how we win. Traction has been very strong and significantly better than our plans, and we are very confident that as the industry grows, it will be an important part of our business. What's interesting is we are now engaging not only in the U.S.; we are also exploring opportunities with hyperscalers and developers in other markets. Our global footprint will help capture that demand more broadly.
Your next question comes from the line of Brian Lee from Goldman Sachs & Co.
I wanted to ask about the battery cell cost uplift. It sounds like it's an international supplier, but can you give us a little bit more detail? Does this have anything to do with the new AESC ownership? Or maybe just walk us through what's changing to impact costs here? And is this also a drag into your fiscal 2028 procurement and cost as well?
Great question. No, this relates to our international supply for international markets and is not connected to AESC. We entered into a long-term agreement that includes supply and technological alignment on how the batteries will work in our modules and how we will work going forward. It's a longer-term contract that we believe will put us in a very good position for 2027 and 2028 and beyond. However, we had to take a charge on one project that was being supplied by that same supplier; as part of finalizing the deal, an adjustment required a charge to that project. The net present value of that project remains significantly higher than the charge we're taking, so we decided to take the charge and move forward. As we continue to grow, integrating our technology roadmap with our cell suppliers' roadmaps is fundamental to our longer-term success.
Fair enough. And then maybe just a question on the conversion cycle because on Slide 7, obviously, a lot of backlog growth in the past couple of years, a lot of pipeline growth, especially the past couple of quarters. So you have a lot of top-of-funnel momentum heading into fiscal '28. Obviously, there are some operational challenges here that are tripping you up in terms of meeting expectations this year. But how should we think about the conversion cycle on these record backlog levels and the impact of these data center and hyperscaler bookings? Just any sense of how quickly we should start to see these turn into P&L impact? And does it differ from your historical backlog conversion cycles?
As you noted, we are just starting with data centers, so we have limited proof points. The proof points we do have show that they can convert much faster. For example, the deal with the developer converted from lead to contract in less than three months, which is tremendously fast. We believe these projects will accelerate our conversion cycle. Our other segments, which represent about 90% today, continue to work on the more traditional conversion cycle of roughly a year to 18 months. Revenue recognition occurs in accordance with project milestones. We typically recognize revenue over an 18-month period, with some revenue recognized earlier as milestones are met during execution, not just at the end.
Your next question comes from the line of Julien Dumoulin-Smith from Jefferies.
A couple of things real quickly, just to rehash. Number one, you made comments about the expedited nature of the behind-the-meter potential customers. Can you elaborate a little bit more about how you're thinking about the potential cadence for incremental bookings from here against some of these arrangements? Also, elaborate a bit on the composition of customers. Last quarter we talked about a couple in particular. Behind-the-meter could be an array of different counterparties as well. Can you talk about the nature of these counterparties? And I got a quick follow-up.
Last time we talked mostly about hyperscalers. We intentionally engage hyperscalers because they often introduce us to developers. With hyperscalers, the procurement process focuses heavily on quality-of-power solutions and involves deep technical evaluation. With developers, the focus is mostly speed to power and faster decision-making. What we have observed so far is developers are in a much greater hurry compared to hyperscalers and tend to convert leads to orders more quickly. The pipeline today still has a majority of opportunities tied to hyperscalers, but the developer segment is the growing portion that is currently active.
Understood. Excellent. And then just quickly, I see a comment here about strategic expenses just of late. How are you thinking about the company strategically, and is there anything to flag regarding quarterly expenses? Is there anything we should watch related to strategic initiatives or costs?
We're always evaluating strategic options and spending time in the market. Earlier this year we explored AESC as an option, and some costs this year are connected to that review and associated analysis. There is nothing to announce at this time. We are actively talking with a range of battery capacity suppliers in the market and assessing opportunities to work more closely with U.S.-based battery manufacturers, but nothing substantive to report yet.
The next question comes from the line of Dylan Nassano from Wolfe Research.
Just wanted to check, so in terms of the scope of the delays that you guys are seeing from the manufacturing facilities, to what extent, if any, are those impacting the hyperscaler MSAs and these orders that you guys are talking about here?
No, not at all. These issues relate to contracts that we signed a year to 1.5 years ago in our traditional segments and are not connected to the hyperscaler MSAs. They do not affect the MSAs or the contracts we're signing with data centers.
Got it. Okay. Maybe it would just be helpful if you could level-set us again on the number of MSAs. I believe it was two last quarter. How many hyperscalers exactly does that include? And can you clarify: you have the $300 million first order and then the $550 million awarded—are those from the same hyperscaler or different hyperscalers?
We have two MSAs with two hyperscalers. As we engage with hyperscalers, they often refer us to developers. The $300 million order was with a developer that was referred by one of the hyperscalers with which we have an MSA; that was a developer building a data center for one of the hyperscalers. The approximately $550 million in awards came from one of the hyperscalers under an MSA and is currently in the tender/award phase; we are finalizing technical points and expect to convert these awards into signed orders in the coming months. We continue to engage with additional hyperscalers, and there are multiple projects both in the U.S. and internationally that we intend to bid on.
Your next question comes from the line of Vikram Bagri from Citi.
I wanted to ask about margins first. One of your larger peers indicated margin pressures in storage. I was wondering how you feel about the 10% to 15% margin guidance. You have backlog converting $2.2 billion next year and $2.8 billion after that. So pretty long-dated backlog at this point. Are there any variances in margins when you look at the backlog in the near term, medium term and long term? Are you witnessing the same pressures your peer talked about?
We remain comfortable within the 10% to 15% adjusted gross margin range. If you adjust for some of the onetime items this year, our margin would have been around the 12% level we guided the market toward. We are confident our backlog and new orders are aligned with the 10% to 15% range. I can't speak to the specifics of our peer's cost structure; there may be company-specific factors. For us, the primary challenges this year are related to scaling—execution and scale-related pains. Scale drives competitiveness and we are focused on resolving those execution issues.
Got it. And on scale and guidance for next fiscal year, how are you thinking about setting guidance? You have $2.2 billion of backlog for fiscal '27. Is 85% coverage the right way to think about it? Or should it be connected to capacity that's coming online, how much you can bring online, and supplier capacity? Should 85% coverage be higher? Relative to that, encouraging to see management changes to improve execution—what specific changes will Roman and Peter make in the next 12 months? You're dealing with contract manufacturers, so they have relatively less control over operations. So what changes will ensure on-time deliveries?
Regarding coverage, we still believe 80% to 90%, with 85% as a reasonable midpoint, is the right way to think about revenue coverage for fiscal '27. We learned from the new facility ramp that even with hedges and contingency plans, unforeseen issues can exceed those contingencies. For next year, we will likely apply stronger hedging around new facilities where appropriate. That said, we do not expect any new major enclosure facilities that would materially change our 2027 revenue assumptions. On the organizational changes: Roman will lead our manufacturing and supply chain given his deep experience in transformation and supply chain operations. As you scale, systems and processes need to evolve; Roman will focus on transforming our processes, systems and planning to operate at the larger scale. We have very good manufacturing partners; this is about execution, not changing suppliers. Peter will continue to lead product development, ensuring Smartstack evolves and remains tightly integrated with our software. The two leaders together will strengthen the connection between hardware and software and drive the operational transformation required for scale.
Coverage.
Your next question comes from the line of Justin Clare from ROTH Capital.
I wanted to ask about the guidance. Based on the revised revenue and adjusted EBITDA guidance, it looks like fiscal Q4 gross margin could be roughly 12% around that range. Can you clarify what's embedded in the assumption for gross margin in Q4? And are there any costs associated with the new products or the production delays from fiscal Q3 expected to extend into fiscal Q4?
Justin, the implied gross margin we're looking at is roughly 11% for Q4 based on the guidance. It's a little lower at the midpoint if you're comparing midpoint to midpoint. Yes, we have considered additional costs that we may incur based on the outlook today. There are many puts and takes, but net-net, after considering those additional costs for delays, we feel pretty good that we'll land in the guidance range we provided based on the current outlook.
Got it. And then on supply chain, could you discuss the potential impact of the FCC's restrictions on inverters? I know you have access to U.S.-based inverter suppliers. Do you have any exposure to sourcing inverters from China? And could compliant domestic sourcing of inverters be a competitive advantage as customers reassess exposure to imported inverters?
We only work with non-Chinese inverters in the U.S., mostly U.S.-made and some imported from Europe, so we do not expect to be affected by the FCC restrictions in the U.S. That has been our policy. We do see a trend toward more regional restrictions and expect Europe will move similarly, and we are preparing for a path to a fully European-compliant solution there. From day one, our view has been that technological and regulatory restrictions will increase as this technology grows in importance to the grid, and we're positioning supply accordingly.
Your next question comes from the line of Christine Cho from Barclays.
I wanted to start with the order intake of $1.44 billion this quarter. The $300 million of that was for the behind-the-meter project and the rest, the $1.14 billion, was more typical front-of-the-meter projects. For the $1.14 billion, how should we think about the split between U.S. and international? And for the $850 million in awards with hyperscalers or data centers collectively, do those include EPC? Is there any difference between developer versus hyperscaler projects? And should we assume that both of these projects have a duration of two hours?
For data center projects, durations are generally two hours; that's a common market standard for these applications. There is no meaningful difference in technical requirements and margins between developer and hyperscaler projects in terms of duration—two hours is typical. Developers tend to be smaller, more agile and make decisions faster, which contributes to quicker conversion. Regarding the $1.14 billion of non-data center orders this quarter, the U.S. continues to be where we are making the most traction; roughly a 60/40 split favoring the U.S. in that quarter. That was a quarter with a lot of U.S. activity and relatively less international activity.
Okay. And then a housekeeping item: the 10-Q indicates there were some IEEPA refunds. Did any of that show up in COGS or was it applied to inventory?
Yes, there is some IEEPA refund—we've recognized a little over $10 million year-to-date. Part of it was recognized in the period and part of it was applied to inventory. As inventory converts to revenue, it will be recognized over time.
Your next question comes from the line of Chris Dendrinos from RBC Capital Markets.
Following up on guidance: the outlook for the remainder of the year has a pretty wide range given the amount of time left. What is driving that guidance range, and how confident are you in the execution path going forward?
That's a fair question. The wider range in guidance reflects prudence. We still have work to do on execution and wanted to provide a range that contemplates potential incremental costs we may incur as operations ramp. The wider EBITDA range is driven by the possibility of additional costs during the ramp-up period. We believe this guidance reflects a prudent view based on current visibility.
Your final question comes from the line of Ameet Thakkar from BMO Capital Markets.
Following up on Chris' question in a different way: if I think about the midpoint of your guidance for fiscal year 2026, it implies about $1.4 billion of revenue for Q4. If I look at revenue recognition and implied ASPs, that implies roughly $0.235 per watt and about 6,000 megawatts of revenue recognition. Are you anticipating a large portion of what you'll recognize in Q4 to include EPC work? Some European contracts historically had attractive implied ASPs because of EPC work— is that the case for Q4? I have one more quick follow-up.
No, most of the expected revenue for the quarter is in the U.S., and that primarily reflects deliveries under domestic content and our typical equipment deliveries. It is not a large EPC-driven quarter.
For the quarter, we have already produced and integrated roughly half of what we need. We are confident in our ability to meet the numbers. The Houston facility ramp is under control; as with any ramp, there are risks we can't fully envision, which is why we kept a wider guidance range. That said, many of the units are already in transit or staged for delivery, so we feel the guidance is appropriate.
Understood. One more: it looks like cumulative deployed megawatts were essentially unchanged versus the prior quarter and up only about 8% or 9% year-to-date. Can you help us understand when those megawatts get reflected as deployed versus revenue? There seems to be a timing disconnect.
Our metric for deployed megawatts refers to projects that have reached substantial completion. Our revenue recognition often occurs earlier—when we deliver equipment to site and transfer title—so revenue is typically recognized before a project reaches our 'substantial completion' metric. That timing difference explains the disconnect. We recognize revenue on a percentage-of-completion basis; 'deployed' as we report it is a different milestone. We may consider aligning our definitions more closely to avoid confusion. Thank you, everybody, for joining. Again, we apologize for the technical issues this morning. It's an inconvenience and we'll ensure it does not happen again. Thank you for your time and your questions.
This concludes today's meeting. You may now disconnect.