Bloom Energy Lands Oracle's Entire 2.45 GW Jupiter Project, Raises 2026 Revenue Guidance 80% at Midpoint to Up to $3.8 Billion
Q1 2026 Earnings Call, April 28, 2026 — Record Revenue, Materially Raised Guidance, and a Landmark Sole-Source Win
Bloom Energy delivered what management called a record first quarter across every major financial metric, and then immediately overshadowed those numbers with a commercial announcement that is likely to redefine how the market prices the company's long-term opportunity. Oracle has selected Bloom as the sole power provider for Project Jupiter, a multi-gigawatt AI data center campus in New Mexico, replacing previously planned gas turbines and diesel backup generators entirely with Bloom Energy Servers in a configuration that will reach up to 2.45 gigawatts. No grid connection. No diesel. No batteries. Just Bloom. The announcement, made by Oracle the evening before the earnings call, is the clearest public validation yet of Bloom's positioning as the primary power infrastructure layer for next-generation AI factories.
Oracle Jupiter: Not a One-Off, But a Template
CEO K.R. Sridhar was emphatic that the Oracle win should not be read as an isolated trophy contract. "Where Oracle is going is where the broader market is headed," he said. More telling is what he disclosed about the composition of Bloom's existing backlog: well more than half of the current data center pipeline comes from customers other than Oracle, spanning additional hyperscalers, neo-clouds, and colocation providers. According to Sridhar, these installations share the same architecture as Jupiter — fully islanded, grid-independent microgrids with no diesel generators, no battery banks for load following, no turbines, just Bloom. The Oracle deal establishes a replicable proof point, but the pipeline it sits within is already substantially larger than one project.
Sridhar framed the win around two converging forces that Oracle explicitly cited when making the switch: community acceptance and speed to power. On the community dimension, conventional gas-fired generation at this scale carries extraordinary environmental consequences. Sridhar offered a striking illustration to put the numbers in context: a 2.5 gigawatt CCGT complex, the conventional alternative, would consume the equivalent of nearly one million showers worth of water per day and generate air pollution approaching that of every car in the state of Rhode Island, concentrated in a single location. Bloom's electrochemical process involves no combustion, uses minimal water at startup and none during normal operations, and operates quietly within a compact footprint. On speed, Sridhar argued that community-friendly permitting translates directly into faster energization, and "at a time where every quarter of delay translates into hundreds of millions in foregone AI revenue and loss of competitive advantage, speed of powered infrastructure development is the difference between leading and following."
Q1 Financials: The Numbers Behind the Narrative
First quarter revenue of $751.1 million represented more than 100% year-over-year growth, the first time in Bloom's history as a public company that quarterly revenue has crossed that threshold. Product revenue reached an all-time high of $653.3 million. Service revenue of $61.9 million grew 15.6% year-over-year. Non-GAAP gross margin expanded 280 basis points to 31.5%, with product margins at 35.3% and services margins at 18%, the latter achieving double-digit profitability for the fourth consecutive quarter. Operating income came in at $129.7 million versus $13.2 million a year ago, a $116.5 million improvement, with operating margins reaching 17.3%, up more than 1,300 basis points. Adjusted EBITDA was $143 million, implying an EBITDA margin of approximately 19%. Non-GAAP fully diluted EPS reached $0.44 versus $0.03 in Q1 2025.
Cash flow from operating activities was an inflow of $73.6 million, positive in the first quarter for the first time, historically the seasonally weakest period. New CFO Simon Edwards, who joined just two weeks before the call, pointed to a step change in profitability, strong collections, and customer prepayments to reserve manufacturing capacity as the key drivers. The balance sheet ended the quarter with $2.52 billion in total cash.
Guidance Raised Materially — Again
Bloom raised its full-year 2026 revenue guidance from a range of $3.1 billion to $3.3 billion to a new range of $3.4 billion to $3.8 billion, with the lower bound of the updated range sitting above the upper bound of the prior range. At the midpoint, this implies approximately 80% year-over-year revenue growth. Non-GAAP gross margin guidance was raised from 30% to approximately 34%, representing a four-point improvement year-over-year and two points above original guidance. Non-GAAP operating income guidance is now $600 million to $750 million, and non-GAAP fully diluted EPS guidance is $1.85 to $2.25. Management guided for Q2 revenue to be at least as strong as Q1's $751.1 million.
The magnitude of the guidance revision is significant. Moving the revenue midpoint from $3.2 billion to $3.6 billion in a single quarter, while simultaneously raising margin targets, signals that both the volume and the economics of the pipeline are improving concurrently rather than trading off against each other. Edwards noted that operating leverage in the model is real: revenue growth is substantially outpacing cost growth.
Manufacturing Strategy: Continuous Expansion, Not Step Functions
One of the more substantive operational disclosures came in response to analyst questioning about how Bloom's approach to capacity expansion has shifted. Sridhar described a pivot away from lumpy, one-time capacity additions toward what he called a continuous analog ramp, adding "hundreds of megawatts a quarter." The current manufacturing footprint supports up to five gigawatts of annual production, and Sridhar said the company intends to reach that level, with additional greenfield factory construction to follow as demand requires.
Critically, Sridhar stated that Bloom is currently neither order-constrained nor capacity-constrained. "The pace of our revenue growth is decided by how fast our customers can build their greenfield sites, not how fast we can power them." This is a meaningful claim. It implies that the binding constraint in Bloom's revenue trajectory is customer construction timelines, not Bloom's own production or order pipeline, which inverts the typical manufacturing bottleneck dynamic and places the ceiling on growth somewhere well above where the company currently sits.
On labor and supply chain, Sridhar made a pointed observation that as production volumes approach ten times the level of a few years ago, the headcount on the factory floor will remain essentially unchanged, a consequence of automation and workforce upskilling rather than linear labor additions. He extended the same philosophy to supply chain partners, describing them as custom-built extensions of Bloom's manufacturing model operating under equivalent efficiency expectations.
Installation Time Reduced by an Order of Magnitude
In response to a question from Colin Rusch of Oppenheimer about installation cadence, Sridhar disclosed an innovation that Bloom has not previously publicized. The company has shifted from a traditional civil construction-heavy installation model — requiring site preparation, conduit laying, and on-site trade labor — to a skid-mounted solution that arrives ready to connect with minimal field work. The result, per Sridhar, is "an order of magnitude reduction in the field time that it takes for us to install our systems." He stated that Bloom can bring a 100-megawatt project online faster and with fewer field hours than any competing technology. For customers for whom time to revenue from GPU compute is measured in hundreds of millions of dollars per quarter, this is not a secondary consideration.
Service Contracts: 100% Attach Rate, 10 to 15 Year Terms
JPMorgan analyst Mark Strouse raised a question relevant to the quality of Bloom's revenue model, asking whether service contract durations are extending as data center customers come on board. Sridhar confirmed a 100% service attach rate across all product sales, including data center deployments, and said the average contract duration for hyperscale customers runs 10 to 15 years. This is a structural annuity embedded in every product sale and is already contributing to the services margin expansion visible in the financials. Services margins at 18%, up 13 percentage points from a year ago, suggest the business is achieving genuine scale leverage in the installed base, not merely growing the numerator.
Inference as the Next Growth Layer
Evercore analyst Nick Amicucci raised the question of whether Bloom's current backlog is predominantly tied to AI training workloads and whether inference represents an incremental demand leg. Sridhar confirmed the framing and extended it. "Inference is going to be much bigger than training in terms of total gigawatt need," he said, but noted that inference compute is distributed, closer to population centers, and therefore even more sensitive to community acceptance of the associated power infrastructure. He argued that the regulatory and permitting resistance already emerging around large training campuses in remote areas will be substantially more acute in urban and suburban inference deployments, reinforcing the competitive advantage of Bloom's clean, quiet, compact profile. Sridhar also confirmed that the "bridge power" framing that hyperscalers once used to describe Bloom deployments — implying a temporary solution until grid connection was secured — has effectively disappeared from customer conversations.
DC Architecture and Grid Infrastructure Bottlenecks
UBS analyst Manav Gupta raised a technically detailed question that surfaced one of the more compelling structural arguments for Bloom's positioning. He noted that competing on-site solutions rely heavily on battery packs for load balancing and backup, which are expensive, thermally intensive, and degrade over time. He also highlighted that the transition to 800-volt DC power architectures in next-generation GPU clusters creates downstream demand for large power transformers, medium voltage switchgear, and centralized rectifiers, all of which are currently facing significant procurement backlogs and lead time extensions. Bloom's architecture, combining fuel cells with ultracapacitors and the ability to deliver direct DC output, sidesteps the transformer and rectifier bottlenecks. Sridhar agreed with the characterization and stated that the shift to direct 800-volt DC delivery is "self-evident" and "inevitable" given global copper and transformer constraints. "Once they try it, they will not go back on it."
New CFO, New Balance Sheet Discipline to Be Tested
Simon Edwards, previously CFO at Grok AI and with prior experience scaling manufacturing operations for complex systems, joined Bloom approximately two weeks before the earnings call. His prepared remarks were measured and credible, emphasizing operating leverage, the strength of the demand pipeline, and a bias toward converting demand into cash flow. He declined to offer long-term financial guidance, noting that executing on the near-term growth vectors is the immediate priority. It will take several quarters to assess his impact on financial discipline and capital allocation, but his background in scaling software and AI-adjacent businesses is at minimum a reasonable fit for the trajectory Bloom is on.
Geographic and Vertical Concentration Remain Real Risks
Sridhar was candid that international demand, while a long-term opportunity, lags the U.S. materially, driven in part by the energy policy disruptions in Europe and slower AI infrastructure development outside North America. On an "80-20 rule," he said, the action today is in the U.S. He also acknowledged that while commercial and industrial demand continues to grow and represents a meaningful part of the business, the size and velocity of AI data center orders have consumed most of the commercial attention. Bloom's revenue trajectory is therefore heavily concentrated in a single vertical and a single geography at this moment, even if both the vertical and the pipeline are large. Any deceleration in U.S. AI infrastructure spending, whether from regulatory, macroeconomic, or technology shifts, would have an outsized impact on Bloom's near-term numbers.
Bloom Energy Deep Dive
Business Model and Monetization
Bloom Energy operates at the vanguard of distributed power generation, having systematically transitioned from a niche provider of behind-the-meter backup solutions into a primary, bankable power provider for gigawatt-scale infrastructure. The core economic engine of the company is driven by the manufacturing, sale, and installation of its proprietary Solid Oxide Fuel Cell (SOFC) Energy Servers. Unlike intermittent renewable assets, these modular servers provide continuous baseload electricity through an electrochemical process without combustion. Bloom monetizes this hardware through upfront equipment sales paired with highly lucrative long-term service agreements. This dual-pronged model ensures a steady baseline of recurring revenue. Indeed, the service division has transitioned from a loss leader into a structural profit center, generating six consecutive quarters of non-GAAP profitability by early 2026. Furthermore, Bloom has specifically tailored its newest server iterations for the artificial intelligence era, offering 800-volt DC-ready systems that plug directly into AI server racks, eliminating the 10% to 15% energy loss typically associated with traditional AC-to-DC data center power conversions.
Customers, Competitors, and Market Position
The company's customer base has rapidly evolved from conventional commercial and industrial facilities to top-tier hyperscalers, global infrastructure asset managers, and utility-scale developers. Over the past year, Bloom has secured a series of transformative, industry-defining commitments. Most notably, the company executed a master agreement with Oracle to deploy up to 2.8 gigawatts of fuel cell capacity, with an initial 1.2 gigawatts actively deploying across United States compute projects. Concurrently, an infrastructure partnership with Brookfield has committed $5 billion to deploy Bloom fuel cells across global AI factories, while American Electric Power signed a $2.65 billion unconditional purchase for 900 megawatts at a Wyoming data campus. Internationally, South Korean conglomerate SK ecoplant remains a critical foundational partner, routinely executing multi-hundred-megawatt deployment contracts. In the competitive arena, Bloom contends with legacy gas turbine manufacturers, traditional grid utilities, and alternative stationary fuel cell providers such as FuelCell Energy, Plug Power, and Doosan-HyAxiom. The market structure, however, remains heavily consolidated. Bloom stands as the unquestioned titan of the space, commanding an estimated 44% global market share among the top stationary fuel cell providers and holding the dominant monopolistic position in the North American commercial and industrial solid oxide sector.
Competitive Advantages
The company's most formidable competitive moat is fundamentally temporal: deployment velocity. In crucial United States power markets, grid interconnection queues for 100-megawatt data center campuses now stretch between seven and ten years. Bloom circumvents this crippling supply chain bottleneck entirely by deploying on-site generation in as little as 55 to 90 days. Beyond speed-to-market, Bloom's technological advantage is deeply rooted in physics. Operating at exceptionally high temperatures, solid oxide technology achieves 52% to 65% lower heating value electrical efficiency in standalone configurations, and scales up to 85% in combined heat and power setups. This represents a 20% to 30% efficiency premium over traditional combustion gas turbines. Additionally, unlike proton-exchange membrane fuel cells, Bloom's high-temperature solid oxide chemistry does not require expensive precious metal catalysts like platinum, shielding the corporate supply chain from acute commodity shocks. The server architecture is also inherently fuel-flexible. It is capable of operating on natural gas, renewable natural gas, or a pure hydrogen blend, effectively future-proofing the massive installed base against evolving environmental regulations and preventing stranded assets.
Industry Dynamics: Opportunities and Threats
The macroeconomic tailwinds propelling Bloom are inextricably linked to the structural power constraints of the global artificial intelligence boom. AI training clusters require hyper-dense power loads that regional utilities simply cannot provision on a commercially viable timeline. Goldman Sachs projects that 8 to 20 gigawatts of fuel cell capacity will be required to supply data center electricity by 2030, presenting a generational opportunity for decentralized power solutions that bypass the grid. Conversely, the primary threats to Bloom's trajectory center on transitional fuel economics and the inherent capital intensity of scaling industrial manufacturing. While the energy servers are structurally designed to transition to green hydrogen, the near-term reality is that the vast majority of Bloom deployments currently utilize natural gas. This leaves the operational model exposed to fossil fuel pricing volatility and potential regulatory friction in jurisdictions aggressively implementing carbon taxation. Furthermore, managing the immense capital intensity required to rapidly scale global manufacturing capacity from 1 gigawatt to 2 gigawatts by the end of 2026 presents ongoing execution risks, demanding flawless supply chain orchestration for specialized ceramic materials and advanced alloys.
Future Growth Drivers: Hydrogen and Electrolyzers
Looking beyond natural gas-powered servers, Bloom is aggressively commercializing its Solid Oxide Electrolyzer Cell (SOEC) technology to capture a dominant share of the nascent green hydrogen production market. By essentially reversing the fuel cell process, Bloom's electrolyzers use electricity and water to produce clean hydrogen. Because the solid oxide platform operates at elevated temperatures, it requires significantly less electrical input to split water molecules compared to legacy low-temperature proton-exchange membrane or alkaline systems. When paired with industrial waste heat, Bloom's electrolyzers achieve an unparalleled 80% electrical-to-hydrogen efficiency, generating 20% to 25% more hydrogen per megawatt of input. This technological superiority is currently being validated through large-scale global pilot projects, including a 1.8-megawatt demonstration deployment on Jeju Island, South Korea, alongside SK ecoplant, and extensive decarbonization studies with energy heavyweights like Shell. As multi-billion-dollar governmental hydrogen hubs mature over the coming years, this electrolyzer division is structurally poised to transition from an auxiliary growth narrative into a primary, high-margin revenue engine.
New Entrants and Disruptive Threats
The stationary power and fuel cell sector presents exceptionally high barriers to entry, dictated by immense upfront capital requirements, decades of proprietary materials science research, and the necessity of proven field reliability. Consequently, there are very few credible new entrants possessing disruptive solid oxide technology capable of matching Bloom's scale. However, peripheral threats exist at the edges of the ecosystem. European players such as PowerCell Sweden have recently begun field validating alternative hydrogen-powered systems for niche data center operators. Additionally, advancements in utility-scale battery energy storage systems integrated with solar or wind assets present a competing vision for off-grid power generation. Yet, the distinct requirement of modern AI data centers is uninterrupted, absolutely reliable baseload power. Renewable-plus-storage configurations simply cannot provide 100-megawatt continuous loads economically across multi-day weather lulls. While small modular nuclear reactors represent a theoretical existential threat to fuel cells for baseload data center power, their commercialization and regulatory timelines remain far too distant to disrupt Bloom's immediate multi-gigawatt pipeline through the end of the decade.
Management Track Record
Under the leadership of Chief Executive Officer K.R. Sridhar, Bloom Energy's management team has successfully navigated the highly treacherous transition from a cash-burning clean-technology pioneer to a sustainably profitable industrial manufacturer. The executive execution over the last two years has been clinically precise, validating the strategic pivot toward the hyperscale data center market. Management correctly anticipated the utility grid bottleneck, positioning Bloom as the definitive bridge technology for the global AI infrastructure buildout. This foresight has resulted in a spectacular acceleration of the business, culminating in Q1 2026 revenue of $751.1 million, a 130% year-over-year expansion. More impressively, management has proven capable of driving profound operational leverage on a larger revenue base. By level-loading manufacturing and engineering out structural unit costs, the company expanded non-GAAP gross margins to 31.5% in the first quarter of 2026, while driving non-GAAP operating income to nearly $130 million. By prudently refinancing convertible debt well ahead of maturity and systematically raising forward revenue guidance to between $3.4 billion and $3.8 billion for 2026, the executive team has demonstrated a highly sophisticated grasp of capital allocation, operational scaling, and shareholder value creation.
The Scorecard
Bloom Energy has structurally repositioned itself from a niche clean-technology vendor into a mission-critical infrastructure provider for the global artificial intelligence buildout. The company's solid oxide fuel cell architecture directly solves the most acute bottleneck facing modern hyperscalers: the severe temporal lag in utility grid interconnections. By offering modular, fuel-flexible baseload power that can be deployed in a fraction of the time required for traditional grid upgrades, Bloom commands significant pricing power and deep strategic integration with the world's largest data center developers. This dominant market positioning is fundamentally validated by a massive $20 billion forward backlog, a staggering 130% top-line expansion in early 2026, and expanding non-GAAP gross margins that underscore the inherent operational leverage of the business model.
Despite this formidable commercial momentum, the long-term thesis rests heavily on managing the bridge between current natural gas dependency and the broader realization of the green hydrogen economy. While the immediate demand profile is highly de-risked by multi-gigawatt master agreements with top-tier technology and infrastructure firms, the company faces rigorous operational parameters as it aggressively scales manufacturing capacity to 2 gigawatts by the end of the year. Furthermore, maintaining its technological supremacy in solid oxide electrolyzers will dictate its ultimate ability to capture the green hydrogen total addressable market. Ultimately, Bloom possesses a rare combination of proprietary physics, unassailable deployment velocity, and demonstrated manufacturing scale, establishing it as the preeminent physical asset in the global transition to decentralized, high-density power networks.