Bloom Energy Q3 revenue up 138% to $1.24 billion, AI power shortage drives demand for on-site power generation

The global AI wave has not only ignited demand for computing power, but has also completely set ablaze a trillion-dollar market for "on-site power."

As tech giants' hunger for data center power reaches a fever pitch, "energy storage meme stock" and fuel cell technology company Bloom Energy has delivered a third-quarter report that can only be called stunning. On October 22, the newly released Q3 financial results showed that, driven by strong demand for on-site power solutions, the company's Q3 revenue doubled and soared 138% to $1.24 billion, quarterly operating profit surged more than 500%, and it raised its full-year guidance for the third time this year. Q3 free cash flow was a net outflow of $181 million; the CRO admitted that there is a timing mismatch between large deliveries and cash collection, and warned that capacity expansion costs will weigh on the 2027 gross margin.

The company's CEO said bluntly on the earnings call that the aging traditional power grid can no longer match the speed of AI expansion, forcing tech giants to shift comprehensively to on-site power. "Today, the world is moving from a century-old centralized power model to a new architecture that includes firm, distributed on-site power generation. Bloom is leading this transformation."

Despite strong fundamentals, amid the combined tug-of-war of the broader market and earlier gains, its stock price fell 7% to around $200 in after-hours trading on Thursday.

Blowout Results: Quarterly Profit Exceeds All of Last Year, Guidance Raised Three Times This Year

On October 22, Bloom Energy held its Q3 earnings call. Over the past year, market attention has expanded from AI chips alone to the infrastructure that powers them.

Thanks to strong product deliveries, Bloom Energy achieved third-quarter revenue of $1.24 billion, surging 138% from $519 million in the same period last year and far exceeding analysts' expectations of $1.07 billion. **Among this, core product revenue exceeded the $1 billion mark in a single quarter for the first time (reaching $107 million), up 179% year over year.

The profit side was even more striking. The company achieved third-quarter net income of $250.3 million (GAAP diluted earnings per share of $0.78), compared with a loss of $23.1 million in the same period last year; excluding one-time items, adjusted earnings per share reached $0.94, sharply beating Wall Street expectations of $0.69.

Bloom Energy Chief Financial Officer Simon Edwards emphasized on the earnings call the enormous operating leverage embedded in the company's business model:

"Operating profit for this quarter alone ($286 million) exceeded the total for all of last year. This quarter's operating margin expanded from 8.9% a year earlier to 23.1%, because while revenue grew 138%, operating expenses grew only 26%."

Based on the strong third-quarter performance and clear visibility into the timing of fourth-quarter customer deliveries, Bloom Energy raised its full-year guidance for the third time this year:

Full-year revenue guidance: Raised to $4.4 billion to $4.6 billion (previous guidance was $3.9 billion to $4.2 billion; the midpoint implies year-over-year growth of about 122%).

Adjusted earnings per share (EPS): Raised to $3.10 to $3.40 (previous guidance was $2.55 to $2.85).

It is worth noting that the financial results showed that net cash outflow from operating activities this quarter was $27 million, and free cash flow was a net outflow of $181 million. The CFO admitted that there is a timing mismatch between large deliveries and cash collection, contract assets and accounts receivable continue to accumulate, and as the business scale expands, working capital occupation will further intensify.

At the same time, the company's CFO said that the ramp-up of new production lines and expedited deliveries are pushing up costs in the short term, and multiple cost-reduction projects still need time to materialize; capital expenditure is already four times that of all last year, and the startup costs accompanying scale expansion will put phased pressure on the 2027 gross margin.

In addition, management clearly stated on the earnings call that the pace of quarterly results is determined by "customer readiness and delivery timing nodes," and the time window for a small number of deliveries at year-end will directly determine where full-year results fall within the guidance range, with uncontrollable risk of quarter-to-quarter volatility.

"A Quarter's Delay Costs Tens of Billions": The Speed-Is-Everything "On-Site Power"

Why have tech giants unleashed such astonishing "on-site power" orders? Bloom Energy founder and CEO KR Sridhar gave the most direct answer on the earnings call—time is money.

"Today, the world is shifting from a century-old centralized power model to a new architecture that includes stable, distributed on-site power generation. Bloom is leading this transformation." Sridhar pointed out that the traditional power grid was originally designed to distribute power evenly to a large, dispersed customer base; when faced with the concentrated and enormous power demand of data centers, the grid has been "overwhelmed."

He quoted the exact words of Nvidia's CEO at the GTC conference in Berlin to explain the anxiety of hyperscalers:

"For a gigawatt (GW)-scale data center, every quarter of delay means $10 billion in lost revenue. (For a 100-megawatt data center, a quarter's delay is a $1 billion loss.) This thirst for speed overwhelms any other value proposition in the power module."

In terms of delivery speed, Bloom Energy demonstrated absolute market dominance. Sridhar revealed:

"Just 15 months ago, we announced our first project with Oracle and delivered power to its Utah site in less than two months. We achieved the same goal on our Nebius project in New Jersey—obtaining planning board approval within three months of signing the contract, and just two months after approval, the first block of power was ready for customer use."

Sridhar further emphasized that Bloom is currently the only on-site power solution that simultaneously meets the three major requirements of "fastest approval (no air pollution permit required), fastest deployment, and future scalability." "We will never be the bottleneck for a customer trying to open a data center," Sridhar said confidently.

Imagination Runs Wild: A Trillion-Dollar Carbon Capture Market and a New Ally in Utilities

Beyond the explosive revenue brought by AI today, the market is also highly focused on the sustainability of Bloom Energy's future growth. On the call, management outlined three TAM (total addressable market) expansion directions with enormous imaginative potential:

1. The "trillion-dollar opportunity" of carbon capture and storage (CCS): This quarter, American Electric Power (AEP) announced it will use up to 700 megawatts of Bloom energy servers supporting carbon capture and storage to provide Google with near-net-zero-emissions power—an unprecedented industry mega-deal. Sridhar was extremely bullish on this prospect at the meeting: "Combustion technology currently cannot provide an economical and scalable path to carbon capture, while Bloom's technology is unique in the commercial sector—it can directly provide high-concentration carbon dioxide for capture. When the world is ready for this step, I believe this is a market opportunity worth more than $1 trillion for Bloom."

2. From "grid competitor" to "grid helper" (front-of-the-meter market): In the past, distributed power generation was often seen as a threat to traditional utility companies. But now, facing power shortage crises, utilities are beginning to proactively purchase Bloom's equipment. "Many utilities increasingly recognize that deploying our clean baseload generation equipment in their distribution networks is a faster and more economical way to meet their customers' growing power needs while enhancing distribution network stability," Sridhar noted, adding that the company is achieving shipments in regions with the lowest electricity prices in the U.S., such as Texas, proving the universal applicability of its cost competitiveness.

3. The architectural revolution of 800-volt direct current (800 VDC): Traditional data centers drawing power from the AC grid must go through complex voltage transformation and AC-DC conversion, resulting in enormous energy waste. Sridhar threw out a highly impactful figure: "We showed in a report that in the first five years of operation, a gigawatt-scale data center using an 800-volt DC architecture would save more than $5 billion compared with current industrial-era power supply architectures."

On the capital expenditure side, CFO Simon Edwards emphasized that the company's capital expenditure for the year reached $231 million, four times that of last year, mainly used to expand capacity in Fremont and Newark. But he noted that the company has achieved "self-funding": "While expanding capacity, our cash balance remained essentially flat compared with the beginning of the year (ending cash of $250 million). This margin structure enables us to invest in scaling the business while maintaining a strong balance sheet."

As CEO Sridhar said in his summary: "After 25 years, we are just getting started. Now, our growth is accelerating."

The full transcript of Bloom Energy's fiscal year 2026 third quarter earnings conference call is as follows (AI-assisted translation)

Title: Bloom Energy Third Quarter 2026 Earnings Conference Call

Host: Hello everyone, thank you for joining, and welcome to Bloom Energy's third quarter 2026 earnings conference call. After today's prepared remarks, we will conduct a question-and-answer session. (Host instructions). I will now turn the call over to Michael Tierney; please go ahead.

Michael Tierney, Investor Relations: Thank you, and good afternoon, everyone. Thank you for joining Bloom Energy's third quarter 2026 earnings conference call. In connection with this call, we have filed our third quarter 2026 earnings press release and supplemental financial information with the U.S. Securities and Exchange Commission (SEC) on Form 8-K, and we have posted these materials on our investor relations website; this call will reference these materials throughout. During this call, both in our prepared remarks and in response to questions, we may make forward-looking statements that represent our expectations regarding future events and future financial performance. These include statements about the company's business results, products, markets, customers, strategy, financial condition, liquidity, and full-year 2026 outlook. These statements are predictions based on our expectations, estimates, and assumptions. However, because these statements involve future events, they are subject to numerous known and unknown risks and uncertainties, details of which are discussed in our filings with the SEC, including our most recently filed Forms 10-K and 10-Q. We undertake no obligation to revise any forward-looking statements made on this call. On this call and in the third quarter 2026 earnings press release and supplemental financial information, we will reference both GAAP (U.S. Generally Accepted Accounting Principles) and non-GAAP financial measures. Non-GAAP financial measures are not prepared in accordance with U.S. GAAP and are provided only as a supplement to, and not as a substitute for or superior to, GAAP financial measures. Reconciliations between GAAP and non-GAAP financial measures are included in the materials referenced above and are available on our investor relations website. Joining me on the call today are KR Sridhar, our Founder, Chairman, and Chief Executive Officer, and Simon Edwards, our Chief Financial Officer. KR will first provide an overview of our progress, and then Simon will review the financial highlights for the quarter. After the prepared remarks, we will open the call for questions. I will now turn the call over to KR.

KR Sridhar, Founder, Chairman, and Chief Executive Officer:

Hello, everyone. At the end of September, as we closed another record quarter, I celebrated Bloom's 25th anniversary and the milestone of the company's inclusion in the S&P 500 — I rang the opening bell at the New York Stock Exchange, surrounded by many members of our manufacturing team.

Our mission has not changed in a quarter of a century: to provide clean, reliable, accessible, and affordable power for everyone. After all, in our modern era, electricity is indispensable for economic and physical health, for human dignity and well-being. The last time Bloom rang the opening bell was eight years ago, when we went public. At that time, the market saw us as a niche power provider. Today, we are seen as the benchmark for on-site power and a key force driving the AI revolution. I firmly believe that when we look back on today eight years from now, we will understand that Bloom led the first major architectural transformation in this 150-year-old power industry. Fuel cell-based on-site power will become the preferred method of power delivery across industries and regions, complementing the grid. Countries with inadequate infrastructure will leapfrog directly to distributed on-site power, just as developing countries skipped landlines and went straight to mobile phones.

Hyperscalers were the first to embrace on-site power. They need power quickly and at massive scale. Some individual data centers consume more electricity than some small states. Yet the grid was designed to distribute power evenly among a large number of users spread across a vast area.

The grid can no longer withstand the highly concentrated, massive power demand of data centers. Mandatory cost-plus pricing mechanisms and utility franchise laws further exacerbate the problem. By design, these mechanisms do not incentivize regulated utilities to innovate or improve efficiency. Before the AI era, two decades of stagnant electricity demand led to severe underinvestment, and much-needed upgrades were long delayed.

The speed and scale of the AI revolution have exposed the grid's shortcomings. Hyperscalers scrambled to find any available grid power and quickly realized that building their own on-site power was the way forward. Bloom quickly stood out as the standardized solution. It is community-friendly and safer. Its redundant, fault-tolerant, distributed architecture provides higher availability — always on, resilient to natural disasters, cyberattacks, and physical attacks. It eliminates the need for backup generators. We have advantages in total cost of ownership, safety, accessibility, deployment speed, and sustainability, and we have a clear path to net zero emissions. Now, the same drivers are pushing a large swath of power market users toward on-site power.

According to our recent power survey of more than 200 commercial and industrial enterprises across 15 sub-sectors, one quarter of U.S. commercial and industrial facilities expect to rely entirely on on-site power by 2030; 35% of facilities expect to achieve fully on-site power by 2035. Why? Looking ahead, electrification and digitalization, population growth, and automation mean that factories and businesses will consume large amounts of electricity and place extremely high demands on reliability. In fact, the International Energy Agency (IEA) forecasts that global electricity demand will grow 40% by 2035.

The grid is neither able to expand quickly nor to supply centralized loads with high reliability, just as it struggles to power AI factories. Bloom will become the preferred solution in this massive market segment. We will become the leader in capturing new global demand in the power market, driving the power market to grow from its current size of roughly $3 trillion to roughly $6 trillion over the next decade. So, 25 years in, we are just getting started. And today, our growth is accelerating.

The market's understanding of the value we provide has deepened, and our brand has gained broad recognition. At the start of our Q1 2026 earnings call, we expected revenue to grow 60% year over year. Today, we raised that guidance and expect revenue growth of more than 120% year over year at the midpoint. Our customer pipeline continues to grow, backlog is expanding faster than revenue, and the diversity of customers, verticals, and regions continues to increase.

We are delivering for customers in real terms. Just 15 months ago, we announced our first project with Oracle and delivered power to its Utah site in less than two months. We are achieving the same result with Nebius in New Jersey—within three months of signing the contract, we obtained planning board approval for the fuel cells. Just two months after approval, the first power was ready for customer use. Our business in this important segment is strong and growing rapidly. Communities and policymakers recognize that we are clean and quiet, and that we are the best choice for solid on-site power.

For example, in the 2026 energy plan, Virginia Governor Abigail Spanberger explicitly named non-combustion fuel cells, quoting: "Compared with traditional gas generation, they produce substantially lower local emissions, reducing health impacts on surrounding communities." This is significant because this is not an ordinary jurisdiction—Virginia is home to most of the world's data centers and is at the epicenter of tensions between data centers and communities.

We appreciate Governor Spanberger's recognition of the value our non-combustion solutions bring at a critical moment for the data center industry and the communities that host it. Below are five commercial developments this quarter that fully demonstrate how we are expanding our total addressable market.

First, the U.S. Army conditionally awarded our partner Indelible a long-term lease to build and operate a commercial power generation facility using Bloom Energy servers at a U.S. Army base, with the goal of enhancing power security through our distributed on-site architecture. We see securely powering military and civilian critical infrastructure—such as military bases, water treatment plants, hospitals, and airports—as a growth opportunity.

Second, AEP disclosed that 700 megawatts of Bloom Energy servers equipped with carbon capture and storage (CCS) technology will be used to provide Google with near-net-zero emissions power, the first project of its kind. Unlike turbines and engines, Bloom's technology makes CCS truly feasible at scale. We believe CCS is essential to achieving decarbonization goals and see it as a potential trillion-dollar market opportunity for Bloom.

Third, we previously introduced our expanding footprint in the Rocky Mountains, the Midwest, and the Great Lakes region. Even in states with the lowest U.S. electricity prices, such as Texas, our value proposition is already competitive—and this quarter we are shipping our first units to that state.

Fourth, the utility partnerships we are building are an important channel for moving America forward amid power shortages. A growing number of utilities recognize that deploying Bloom Energy servers—a clean baseload generation asset—on their distribution networks is a faster, more economical way to meet customers' growing power needs and enhance distribution network stability and reliability. I expect we will make significant breakthroughs in our utility-related business (often called "front-of-the-meter" arrangements).

Fifth and finally, we are seeing strong interest in many international markets in deploying Bloom systems to meet rapidly growing power demand. We are actively in discussions with potential local partners in several of these markets and look forward to updating everyone on some good news soon.

In summary, Bloom is the community-friendly on-site power solution of choice—rapidly deployable, clean, reliable, and competitive. Our on-site solutions are deployed to power gigawatt-scale smart factories, megawatt-scale inference data centers, factories, commercial enterprises, and mission-critical infrastructure such as hospitals, telecom hubs, and essential services. Demand for Bloom on-site power continues to grow domestically and internationally, and we are building strong partnerships. This growth potential, combined with future growth opportunities such as CCS and 800 VDC, points to Bloom capturing a significant share of a growing power market. This conviction gives us a clear view of where we are headed.

This will require us to scale and execute effectively to seize this once-in-a-century opportunity. I have confidence in my team and know they are ready to meet this challenge. Next, I will hand it over to Simon to walk you through our financials in detail.

Simon Edwards, Chief Financial Officer:

Thank you, KR. Good afternoon, everyone. As a reminder, I will focus on non-GAAP adjusted metrics. Reconciliations to GAAP are included in our press release and supplemental materials. The third quarter was another record quarter for Bloom.

Revenue reached $1.24 billion, up 138% year over year; operating profit reached $286 million, growing even faster, up more than 500% year over year. With the strong performance this quarter, we raised our full-year revenue, operating profit, and earnings per share guidance for the third time this year. Before diving into the details, it is worth reviewing what we have achieved so far this year. Year-to-date revenue has reached $3.05 billion, already 50% higher than our full-year 2025 revenue.

Operating profit for this quarter alone exceeded last year's full-year operating profit. This is precisely the operating leverage embedded in our business model, now fully reflected in our results. Operating margin expanded to 23.1% this quarter from 8.9% a year ago, driven by revenue growth of 138% while operating expenses grew only 26%. As KR said, to seize this once-in-a-century opportunity, we need both scale expansion and efficient execution, so we will continue to invest in our innovation and engineering roadmap, growth initiatives, and the infrastructure needed to support our scaling.

Year-to-date capital expenditures were $231 million. We are investing in capacity to support future growth and funding these investments through cash flow generated from operations. Even though our investment is 4x our full-year 2025 level, our cash balance is roughly flat versus the beginning of the year. Our profit structure enables us to invest in scaling the business while maintaining a strong balance sheet. This is precisely the financial model we carefully designed and implemented, and it is performing as expected.

Now let me turn to the specifics of this quarter. As I mentioned earlier, revenue was $1.24 billion this quarter, up 138% year over year and up 16% quarter over quarter. Product revenue was $1.07 billion, up 179% year over year and up 15% quarter over quarter, surpassing the $1 billion mark in a single quarter for the first time and accounting for 87% of total revenue. Product revenue is recognized when control transfers to the customer, so it directly reflects system deliveries and acceptances in the quarter.

Service revenue was $67.6 million, up 15% year over year and roughly flat quarter over quarter. Installation revenue was $84.9 million, up 29% year over year and up 66% quarter over quarter. Installation scope varies by customer and project, and revenue is recognized proportionally as engineering progresses. Therefore, installation revenue and product revenue do not move in sync within the same quarter; installation revenue will fluctuate with the timing and mix of customer projects.

On margins, product gross margin was 38.4%, up 250 basis points year over year and up 120 basis points quarter over quarter—achieved in a quarter in which we were simultaneously ramping new production lines and expediting deliveries. This directly reflects our pricing power and the continued cost reduction results we are driving on the product side. Service gross margin was 21.4%, up 700 basis points year over year, marking our second consecutive quarter above 20%. Total gross margin was 34.5%, up 412 basis points year over year, with expansion in both products and services, partially offset by timing factors in the revenue mix from installation revenue. Given this quarter-to-quarter structural volatility, this is also why we manage and guide gross margin on a full-year basis. Operating expenses were $141 million this quarter, reflecting continued investment in our engineering roadmap, growth initiatives, and infrastructure. Operating profit was $286 million, up more than 500% year over year and up 19% quarter over quarter, with an operating margin of 23.1%, up 60 basis points quarter over quarter.

Adjusted EBITDA was $299 million, or 24% of revenue, expanding 1,286 basis points year over year. Non-GAAP diluted earnings per share were $0.94, compared with $0.15 in the prior-year period. GAAP diluted earnings per share were $0.78, compared with a loss of $0.10 in the prior-year period.

On cash, cash flow from operations was a net use of $27 million this quarter. Net income was offset by an increase in working capital, driven primarily by inventory, contract assets, and accounts receivable, as we supported significant business growth and built inventory reserves for future deliveries. As I mentioned earlier, product revenue is recognized when control transfers to the customer. Contract billing follows contractual milestones, which may occur in different quarters. Therefore, cash collection may lag revenue recognition, and the difference between the two is reflected as contract assets and accounts receivable until invoicing and collection are completed.

The timing of large deliveries and related cash collections will vary by quarter, and these differences will even out over time. We expect to continue investing in working capital to support future growth and scale alongside our customers. Capital expenditures were $153 million in the quarter and $231 million year to date, primarily for investment in machinery and production capacity to support our expansion within our existing facilities in Fremont and Newark. Adding capacity in this incremental manner means our capital expenditures track the demand we see rather than running ahead of it. Free cash flow was a net use of $181 million in the quarter and positive $42 million year to date, and our cash balance at quarter end was $2.5 billion. In addition, Bloom was added to the S&P 500 index during the quarter. We view this inclusion as recognition of what we have achieved in terms of business scale, profitability, and stability, and we welcome the broader shareholder base it brings.

On guidance, we are raising our full-year revenue outlook to $4.4 billion to $4.6 billion, with the midpoint corresponding to approximately 122% year-over-year growth. This increase reflects our third-quarter performance and clearer visibility into customer delivery timing in the fourth quarter. Our quarterly cadence is determined by customer readiness and delivery timing, which is why we manage the business and provide guidance on a full-year basis. As with each of our increases this year, our outlook is built on two layers. The first layer is backlog, for which we assess each customer's site readiness and conversion timing individually. The second layer is a probability-weighted assessment of customers in our pipeline who need power supply in the current quarter. As we enter the final quarter, the timing of a small number of deliveries around year-end will determine where we ultimately land within the range.

We are maintaining our latest non-GAAP gross margin expectation of approximately 34%. This is above 30% in 2025 and above our original expectation of approximately 32% at the start of the year. This expectation reflects our forecast for fourth-quarter revenue mix and margins across our business lines. We are also raising our non-GAAP operating profit outlook to $960 million to $1.03 billion. The midpoint corresponds to a 22% operating margin, compared with the 14% operating margin at the midpoint of our February expectation. Our non-GAAP diluted earnings per share expectation is now $3.10 to $3.40.

In summary, we are delivering record results, raising our outlook, and making the necessary investments to capture the enormous market opportunity in front of us. At the same time, we are continuing to expand margins and maintain a strong balance sheet. With that, operator, we are now ready to take questions.

Question-and-Answer Session

Operator: The question-and-answer session begins now. (Operator instructions) Please hold while we compile the question roster. Our first question comes from Mark Strouse of JPMorgan. Your line is open, please go ahead.

Mark Strouse: Okay, good afternoon. Thank you very much for taking our questions. KR, I'm curious—could you talk about the reasoning behind the recent decision to expand your real estate footprint in Fremont? You spent considerable time on this call discussing multiple paths to TAM expansion. What I'd like to understand is why make this decision now? And since you previously mentioned that current manufacturing capacity can support more than 5 gigawatts of output per year, once the new facility is operational, should we think of that number as exceeding 10 gigawatts?

KR Sridhar (Founder, Chairman, and Chief Executive Officer): Mark, thank you for your question. Look, some things are very clear to us—in a world full of uncertainty, some things can be completely certain. Electricity use and electricity demand will rise sharply, not only because of AI, but because in this digital age, far too many market segments need electricity. This is not only happening within the United States; we are already beginning to see the same trend in Asia, Europe, and the rest of the world. We see demand for on-site self-generation, as opposed to grid-supplied power, emerging from multiple areas beyond data centers—data centers that are being asked to solve their own power supply. This demand comes from the pursuit of resilience, from the need for security. Amid the various complex situations around the world today, the need for reliable, secure on-site power for critical facilities is becoming increasingly important. Taking all of this into account, it is very obvious to us, and has always been obvious—in terms of timing, the 5-gigawatt milestone is just one point on a long journey, we have clearly seen it approaching, and we need to prepare for the phase beyond this milestone. That is what you are seeing now. We will never expand capacity in a jump-like manner; instead, we will follow demand growth gradually, like turning an analog knob. We will not go beyond our line of sight, but we will also never be so slow as to become a bottleneck for customers seeking power. That is how we proceed. We will not provide a specific timetable, but we have strong confidence that the signals the market is sending us are clear enough to tell us we should begin preparing for beyond the 5-gigawatt milestone. That is why we are doing this. I hope that answers your question.

Mark Strouse: Very good, thank you. Next is a brief follow-up for Simon. Regarding the fact that the full-year gross margin outlook is maintained unchanged while the revenue expectation has been raised, I understand this is at least partly due to a change in product mix. Could you provide more details? Also, I know you have not yet given 2027 guidance, but directionally, as revenue scale expands, should we continue to expect gross margin to improve further? Thank you.

Simon Edwards (Chief Financial Officer): Okay, Mark, hello. First, I would like to step back and look at how far we have come — overall gross margin was 28.7% in 2024, 30.3% in 2025, and had reached 34.5% in the third quarter. I think we have demonstrated a good track record of continuously improving margins as the business scales, and I am very satisfied with how product gross margins have performed. As I mentioned in my prepared remarks, pricing remains strong, and the cost-reduction initiatives we have been advancing are already reflected in the numbers. But at the same time, we are also balancing margin expansion against capturing the market opportunities in front of us. As Kara mentioned, we are accelerating the ramp of new production lines and expediting deliveries to meet customers' timeline requirements, which will bring certain costs in the short term. In addition, we have multiple cost-reduction projects in engineering, procurement, and manufacturing, and these will take some time to show up on the cost side. So, when you ask about next year, we are considering it from a long-term perspective. We understand that scale expansion comes with startup costs, and I think we have prudently reflected this in the gross margin outlook. But as these projects gradually move into mass production, and as we continue to optimize line maturity and improve capacity utilization, we believe we will gain more benefits from economies of scale. I just want to say that in my first six months on the job, I have been impressed by the team's operational performance and by how these advances are reflected in the numbers.

Mark Strouse: Thank you.

Operator: The next question comes from Nick Amakuchi of Evercore ISI, please go ahead.

Nick Amakuchi: Hi, everyone, thank you very much. Simon, I would like to dig into the free cash flow situation. This quarter clearly used $27 million of cash, and given margin expansion and other factors, I would like to understand this in more depth. What factors or one-time items can help us understand this?

Simon Edwards, Chief Financial Officer: Okay, Nick, hello. First, I want to say that we really view this business as a self-funding business. That is, when you consider the operating performance of the business, it is that performance that is funding the investments needed for growth. I want to emphasize that this is very important to us, because it will reduce our dependence on external capital and truly lower our execution risk. At the same time, I think it also gives us the flexibility that KR mentioned, allowing us to invest in response to customer demand. So I would say that, in this context, cash generation should be viewed on a full-year basis rather than inferred from any single quarter. We do believe that this business will become a free-cash-flow-positive business, because we are able to continuously fund our own growth. Coming back to the first part of your question, we used cash in operations this quarter, but this is not the start of a trend. Even as we increased operating investment in people and capabilities, we have generated positive cash flow from operations (CFOA) year to date. As I mentioned in my prepared remarks, we also funded the working capital (inaudible) deployed for these customers. So I think our ability to support these investments from an operating perspective while also funding year-to-date capital expenditures precisely demonstrates the capital efficiency of this business. When I connect all of this to the returns we get from our capital expenditure investments, we are very optimistic about the trajectory we are on.

Nick Amakuchi: Great, thank you. KR, if I may, I would also like to dig in a bit more. You mentioned earlier the growing demand and relationships in utilities. Now people clearly see you more as a behind-the-meter solution, and from that perspective it seems more complementary. If we consider what has already been made public, we have seen fuel cells included in integrated resource plans (IRPs) in different regions. Could you elaborate more specifically on the market opportunity and total addressable market (TAM) you are seeing, especially with respect to regulated utility opportunities?

KR Sridhar, Founder, Chairman and CEO: Nick, that's a great question. Let's first talk about it from the customer's perspective. Here, I would say, let's consider two types of customers: one is the ratepayer, that is, the end customer; the other is the utility that serves these ratepayers. If you look at it from the perspective of optimizing for these two types of customers, they are increasingly starting to view us as a distribution asset through which they can lower total costs while continuing to serve customers. The better this is for ratepayers, the easier it is for utilities to ask their policymakers to include the procurement of Bloom Energy servers in the rate base and deploy them as transmission and distribution assets into the T&D network. What does that mean for them? It allows them to deploy us in front of the meter, distributing as load arrives, without waiting for a specific customer order. They can procure based on plans and demand, as they do now in integrated resource planning, without needing to target a specific customer. Why would they do this instead of buying centralized assets? Transmission and distribution and their associated costs, bottlenecks and timing issues simply don't work for them. They are recognizing this. If customers have to turn to behind-the-meter solutions, they lose customers and lose rate base. If they lose customers and rate base, they have to raise costs for other ratepayers, which is not what state governments want, nor what end customers want. So, at this point in time, we are a truly valuable asset for the entire industry. As more regions begin to understand this, you will see this unfold gradually across multiple regions. Our team has done an excellent job of promoting to major utilities the benefits of on-site power as a supplement to the grid. We have been explaining why the grid falls short. Utilities have finally gained the room they need from regulators to include these assets in planning, so they are seizing this opportunity. We see this as a huge business growth opportunity in the future.

Nick Amakuchi: Thank you.

David Arcaro: Oh, hello. Thank you very much for taking our questions. AR, I think you have several future opportunities in carbon capture, combined heat and power, and the industry's transition to 800-volt DC, and I would like to understand how these future initiatives are progressing in terms of commercial opportunities, and what market adoption you are seeing in the near term?

KR Sridhar, Founder, Chairman and CEO: David, hello. Thank you for the question.

Okay, let me answer that. You know, I am incredibly excited about the 700-megawatt carbon capture project announced by AEP—a project that will provide near-net-zero emissions power to a hyperscale customer, in this case Google, as they announced. This project will provide timely, on-site, reliable, air-pollution-free, water-consumption-free near-net-zero emissions power. How much better could it get? Okay.

We have—here is a key point that I think investors on the call need to understand. Combustion technology currently cannot provide an affordable and scalable path to carbon capture. Bloom's technology is unique in the commercial world because it can provide concentrated CO2 for carbon capture. That is Bloom's genius and strength. It is good for the wallet and good for the environment. So I believe that when the world is ready to decarbonize, the fastest way is to use our technology with natural gas to produce large amounts of power that is both reliable and friendly to the planet. When the world is ready for that, I think this will be a more than $1 trillion opportunity. That is how big I think it is.

800-volt DC is the obvious choice. In a report we published, we showed that in its first five years of operation, a gigawatt-scale data center adopting 800-volt DC will save more than $5 billion compared with the current outdated industrial-era power transmission architecture. 800-volt DC is coming regardless. We don't need to push the market to make this happen, but why use all the intermediate conversion steps and pay $5 billion more? I am extremely confident that when server racks with 800-volt DC come out, we will see this adoption trend happen.

We are in conversations with all the hyperscalers. There is a lot of interest in this area. It will all happen when DC data centers built with 800-volt DC as the standard are completed. It is not a question of whether it will happen, but how fast, so they can answer the timing question better than I can.

On heat recovery, that is, combined heat and power (CHP), we firmly believe that when you generate power on-site, wasting the heat that could be used for cooling and thereby reduce electricity demand by 20% is likewise something that obviously should not be done. So this is complementary to everything we do, whether you are supplying AC power to a data center or 800-volt DC, with or without carbon capture, it doesn't matter. It becomes an obvious choice. One interesting thing about CHP—most industrial applications, and we believe the commercial and industrial (C&I) sector will be a huge growth direction for us, will need thermal energy in the form of steam or cooling.

They will use combined heat and power. So I think this is a huge growth direction, and also a way to create value by using our products for commercial and industrial customers. The beauty of it is that it is the same platform. That is why we have always told everyone that Bloom is a platform, not a product. We are able to offer them any one or all of the options in the platform, according to what the customer needs.

Thank you.

Operator: Next question.

KR Sridhar, Founder, Chairman and CEO: David may have a follow-up question.

Operator: The next question comes from Chris Dendrinos of RBC Capital Markets. Please go ahead.

Christopher Dendrinos: Okay, thank you very much for taking the question. Congratulations on a strong quarter. I wanted to ask about "time to power"—I recall you mentioned in your earlier remarks that about a year ago you were able to complete delivery for Oracle in less than two months. As you know, demand is now growing rapidly and you are continuing to scale, so can "time to power" still be kept within 90 days, or has it been extended due to the large order backlog? Also, could you be more specific about how quickly you are powering the Nebius site? You mentioned that the first block is about to be powered, but how many megawatts does that specifically correspond to? Thank you.

KR Sridhar, Founder, Chairman and CEO: Thank you, Chris, that is a very good question. Our servers are ready for the Nebius data center. We reiterate that we have indeed fulfilled our commitment to them. We will not be the bottleneck for bringing their entire 300-megawatt data center online. As soon as we are ready, we will move forward immediately.

However, I think you also understand how data centers are built—they do not go from 0 to 300 megawatts in a day, but advance block by block. Their first block is about to be ready, and we are already ready for the first block. Before each of their subsequent blocks is completed, we will be in place ahead of time to ensure that we are not their bottleneck. As for the timing of the rest of the data center, the customer needs to explain that themselves; it is not appropriate for us to comment.

But back to your broader question. Just this morning—I think it was Berlin time, or perhaps yesterday, I don't remember clearly—at GTC in Berlin, NVIDIA's CEO said this: every quarter of delay for a gigawatt-scale data center means $10 billion in lost revenue. That is, a quarter of delay for a 100-megawatt data center means $1 billion in lost revenue. That number far exceeds any other value proposition you can get from power modules. And there is currently no technology cleaner than ours that can obtain air emission permits while generating power on-site; nor is there any technology that can deliver power to customers faster than we can; not to mention the comprehensive future-proofing capability we discussed in the previous question, and all the added value we can provide. That is why we are optimistic about this particular area and what it means for us. Thank you.

Christopher Dendrinos: Thank you. Just a follow-up, going back to Mark's question about expanding the business footprint—is there an opportunity to introduce more automation in this regard? How do you view the manufacturing opportunity, and what cost savings might this bring? Thank you.

KR Sridhar, Founder, Chairman and CEO: I think anyone who has visited our factory will remember a saying from our excellent general manager, Brad Fields. He reminds everyone: years ago, when we produced 100 megawatts of stacks per year, we had about 600 employees working around the clock; and when we reached 1 gigawatt of annual capacity, the entire production line still needed only 600 employees—meaning a 10-fold efficiency improvement. When we reach 5 gigawatts of capacity, that production line will still need only 600 employees.

Bloom is a company with cost reduction and efficiency improvement written into its DNA, and continuous innovation is our instinct. So the answer is yes, we are not only beneficiaries of AI—they are our customers, and at the same time we also benefit from AI in many different areas, including automation.

Christopher Dendrinos: Thank you.

KR Sridhar, Founder, Chairman and CEO: Thank you.

Operator: The next question comes from Manav Gupta of UBS. Please go ahead.

Manav Gupta: KR, I want to go back to the last conference call, where you said some very insightful things. You mentioned that LCOE (levelized cost of energy) is an absurd construct for on-site power and data centers, and that we should focus on total cost. Can you help us understand the difference between the LCOE that data centers see and the actual price of electricity delivered to them, and how your solution is actually more cost-effective? Also, regarding the report you mentioned on September 15 and 16, that was a very good report because it detailed all the money that can be saved by switching to your solution. I remember you mentioned that a 1 GW facility can save up to 9% of costs by switching to your solution. And one last question, I promise this is the last one—we understand that your solution has lower NOx and SOx emissions, but we also think there is a water-saving advantage here. Your solution uses less water than competitors, which could be a major advantage in the competition. If you could address some of these questions, thank you very much.

KR Sridhar, Founder, Chairman and CEO: Okay, Manav, thank you. There are a lot of questions, and I will try to remember them and answer them one by one. Let me start with the water issue.

You see, if you compare the water usage of a power plant per gigawatt of data center between combustion technology and our technology, our water usage does not even amount to a rounding error at the second decimal place. That is, if you operate a gigawatt-scale data center using combined cycle gas turbines (CCGT), the annual water usage over the entire life cycle is extremely large—for example, if wet-cooled CCGT is used, over the entire life cycle of that 1 GW data center, about 25 billion gallons of water would be consumed, which is a very large number.

Compared with those 25 billion gallons, we need only 1.4 billion gallons. From a water usage perspective, that is a drop in the ocean. Okay, that is how to understand the water issue.

Now, you asked about the cost advantage compared with grid electricity.

From grid electricity to rack power, electricity goes through layer upon layer of transmission, from three nines of reliability (99.9%) to instability at the voltage and power level—I am not talking about environmental cleanliness, but about the cleanliness of power quality, which is exactly what UPS (uninterruptible power supply) is meant to solve, plus reliability issues. Add to that the process of converting electricity to the voltage and direct current you need, and these are all costs brought about from grid electricity to rack power.

In addition, inside the rack, because technologies delivered by the grid cannot ramp load up and down quickly, and gas turbines also cannot ramp load up and down quickly, inefficiencies arise inside the rack. To avoid burning out the turbine, engine or battery on the other side, they can only burn off that electricity for nothing. With fuel cells, there is no need to do that at all.

Therefore, for data centers, the real cost is the ratio between the total cost of electricity and the actual compute output (tokens), that is, the electricity cost per token. From LCOE, to the cost of electricity delivered to the grid, to the cost of rack power, and finally to the electricity cost per token. Adding all of these together, we are the overwhelming cost leader and the lowest-total-cost solution. This is exactly what people are beginning to understand.

I hope I have answered most of your questions, Manav. Let us move on to the next question.

Manav Gupta: Thank you very much, thank you very much, sir. Thank you.

Operator: The next question comes from Ben Kallo of Baird. To save time, please limit yourselves to one question each, all remaining analysts, thank you. Ben, your line is now open, please go ahead.

Ben Kallo: Hi, thank you. I have two questions, quickly. First, going back to Mark and the topic of expansion, you previously talked about expanding by, I recall, 250 megawatts to 300 megawatts per quarter. I just want to understand, as you approach 5 gigawatts next year, how we should think about that, and whether there are any supply-side constraints to consider. The second question is about the "NIMBY effect." We have seen some reports that perhaps some developers are moving overseas. I want to know, if that is true, how we should expect you to benefit from it, and whether your advantages in water, noise, sulfur oxide and nitrogen oxide emissions also apply overseas. Thank you.

KR Sridhar, Founder, Chairman and CEO: Thank you, Ben. On the first question, let me be clear: we have never given any specific numbers on quarterly growth or annual growth. What we have said is that this is like a knob, and we will turn it up or down based on demand. Based on everything I just told you, I don't think we will be turning it down anytime soon; we will only keep turning it up. So how much we turn it up depends entirely on load demand—that is the first point. Second, on the question about international markets, keep in mind that if the total delivered cost of power per token is far lower than the cost of consuming natural gas using a Bloom system, that means less fuel is consumed to produce the same number of tokens. In most international markets, natural gas prices are much higher than in the United States, so the advantage we bring is further amplified. For all these reasons, international markets are very attractive to us, and we are considering international expansion. As I said in my prepared remarks, please expect us to bring you some good news on this front in the near future.

Ben Kallo: Thank you.

Operator: The next question comes from Ameet Thakkar of BMO Capital Markets. Your line is now open, please go ahead.

Ameet Thakkar: Congratulations on the quarter's results and the raised outlook. I just have one quick question, and I will follow the rules. KR, you talked a lot today about the question of scale. What is the largest single-site Bloom deployment currently operating in the United States, in megawatts? And how should we think about that compared with the scale of the projects you mentioned are coming in Wyoming or New Jersey?

KR Sridhar, Founder, Chairman and CEO: We are already at the 100-megawatt level, the Wyoming project is also at the 100-megawatt level, and it is moving toward the gigawatt level, because it reached 700 megawatts. So our smallest system is 150 kilowatts, powering a small Home Depot, and our largest system is approaching the gigawatt level. I can tell you that every customer matters to us, and we keep all our customers satisfied. Thank you.

Ameet Thakkar: Thank you.

Operator: The next question comes from Maheep Mandloi of Mizuho. Your line is open, please go ahead.

Maheep Mandloi: Hi, good evening. Thanks for taking the question. This is a small question. On backlog growth, I recall in the prepared remarks you mentioned that backlog is growing faster than revenue. I want to understand how to interpret that—given that revenue grew about 140% year over year. So is that a useful frame of reference for understanding backlog growth this quarter? Thank you.

KR Sridhar, Founder, Chairman and CEO: Maheep, thank you for your question. Fortunately, you won't have to wait long, because it is now the fourth quarter, and it is the start of our year. We will give you specific guidance. I think over the past few quarters we have proven to you that we are disciplined, and that the numbers we give are ones we can actually deliver. So we will give you that number. But I would say growth is solid and strong. So please expect solid growth. I won't give you a range on the direction, sorry.

Unknown Speaker: Okay, thanks for the question.

KR Sridhar, Founder, Chairman and CEO: Yes.

Operator: The next question comes from Sunaina Ocalan of Bernstein. Your line is open, please go ahead.

Sunaina Ocalan: Hi, good afternoon. Thank you for taking my question and giving me the opportunity to ask. Maybe I can ask, given the AT deal with Google, whether you can share some color on your discussions with your other hyperscale data center customers? Are you seeing carbon capture and storage (CCS) increasingly mentioned as a potential option in hyperscale customer discussions? And from your perspective, what impact does that have on lowering the cost of power? I appreciate the point that your solution has cleaner CO₂ emissions. I would love to hear more of your thoughts.

KR Sridhar, Founder, Chairman and CEO: That is a very good question. I would say yes, all of our hyperscale customers are currently showing strong interest in the optionality of carbon capture. We believe this will become extremely important as we move forward. I think almost all of our customers will soon at least begin buying carbon-capture-ready systems from us, so that when a CO₂ pipeline is built in the ecosystem in their region and connected to a sequestration well, they will be able to connect—that is the first point. Second, think of 10% to 15% of the cost as a reasonable cost that most hyperscale customers are willing to accept today in order to achieve zero-carbon power supply. If that can be provided, it will be the direction they voluntarily choose without a mandate, and Bloom will be able to deliver within that cost range, covering the full integrated cost of carbon capture and storage. That is why I say we are the most economically viable and technically mature carbon capture and storage platform.

Sunaina Ocalan: Thank you.

Operator: Your next question comes from Colin Rusch of Oppenheimer. Your line is now open, please go ahead.

Colin Rusch: Thank you very much, everyone. This is a two-part question. The first question is mainly about the growth in contract assets and how you expect to convert them into a billable and collected status. The second question, or the second part, is mainly about the strategy for monetizing carbon capture. I think we believe that you can indeed have a significant impact in this area. Do you expect to monetize through server sales, or might you launch a new solution to expand some of the opportunities already present in the current service offering?

KR Sridhar, Founder, Chairman and CEO: Okay, Colin, I'll answer the first part.

Regarding the characterization of contract assets, I do expect them to turn over. So while the contract asset balance may grow, what really matters to us is that they are continuously turning over—because we are reaching billing milestones and achieving cash collection. I think this is an important detail that I want to emphasize in particular. Overall, I expect that as the business scales, we will be able to manage working capital continuously and effectively. But I think ensuring that both accounts receivable and contract assets continue to turn over is our primary focus.

If the end customer happens to be a natural gas company or an oil and gas company with carbon capture capability, they can attach this feature and monetize it through carbon credits as well as revenue streams from customers. We can also collaborate with partners to provide a complete end-to-end solution and negotiate how to share the proceeds. Or in some cases, the end customer itself may have a need for carbon credits, use them itself, and pay us through an implied cost of electricity or capital equipment plus a margin. We are open to all of the above models.

That is the beauty of our model—we create multiple possibilities for our customers. This is also how we think about all of our business. Ultimately, the reason we exist is to serve customers and meet needs, and there are many paths to that goal.

That said, since we have reached or even exceeded the scheduled time, let me wrap up. I want to say that while everything you have heard from us today is very positive, I also want to acknowledge the turbulent times we are currently in. Like most of you, I see a great deal of uncertainty in the world and volatility in the markets.

However, it is precisely at times like these that we should remind ourselves that there are ample reasons to remain optimistic about our shared future. Personally, I can tell you that I feel very fortunate to be at Bloom. For me, this is an island of stability, certainty and mission. Let me explain.

I grew up in a place where electricity was scarce. I know deeply what it means to be unable to rely on a stable power supply, how it suppresses economic activity and harms human health and well-being—I have experienced all of this firsthand. In today's digital world, having electricity is even more important than it was in my childhood.

There is no doubt that electricity demand will continue to rise, and this trend is cyclical. One thing is certain: in the coming years, demand for electricity will climb sharply. And Bloom has built a platform that can deliver electricity to those who need it in a clean, reliable and large-scale way. As we have repeatedly demonstrated, we can achieve this, create profits and returns for shareholders, and drive scalable growth of the business through a self-funding model.

Our future is bright. I am very grateful to be able to shoulder such a meaningful mission together with such a strong team. It energizes me, and I can clearly see our bright future. Thank you all for your support of Bloom, and have a good day.

Operator: That concludes today's conference call. Thank you for your participation. You may now disconnect. Risk Disclosure and Disclaimer

Markets carry risk, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation or needs of individual users. Users should consider whether any opinions, views or conclusions in this article are suitable for their particular circumstances. Investment based on this is at the user's own risk.