Bloom Energy ($BE) Stock Hub: Record Q2, AI Power Demand, Concentration and the Valuation Test

1. Bloom Energy Investor Snapshot

Bloom Energy designs, manufactures, sells and services onsite power systems based on solid oxide fuel cell technology. Its core product, the Bloom Energy Server, converts fuel into electricity electrochemically rather than through combustion. The company’s most important current use case is reliable, modular power for data centers, semiconductor facilities, utilities and other commercial or industrial customers that cannot wait years for new grid capacity.

The investment story changed dramatically during 2025 and 2026. Bloom moved from a company often grouped with speculative hydrogen and fuel-cell equities into a profitable operating platform at the center of AI infrastructure spending. Full-year 2025 revenue reached $2.02 billion, first-half 2026 revenue reached $1.82 billion, and management raised full-year guidance to $3.9–$4.2 billion after record Q2 results. The company also has major structures with Oracle, Brookfield, AEP, Equinix, CoreWeave and Nebius.

That progress is real. It does not mean every announced gigawatt is a firm backlog item or that every financing framework becomes Bloom revenue. Investors need to separate master agreements, financing capacity, purchase commitments, customer deposits, recognized backlog and actual quarterly deliveries.

2. Company Profile and History

Bloom Energy was founded by KR Sridhar, who serves as Chairman and Chief Executive Officer. The technology traces part of its history to work connected with NASA’s Mars exploration program, where solid oxide electrochemical systems were developed around the challenge of converting gases into useful power and resources. Bloom later commercialized the platform for terrestrial distributed generation.

The company is headquartered in San Jose, California and trades on the New York Stock Exchange under the ticker BE. It employs more than 2,000 people and manufactures primarily in the United States, with additional assembly activity through a Korean joint venture.

Bloom’s history contains long periods of high expectations, losses and uneven margins. The company went public in 2018 and spent years proving that the Energy Server could become more than a niche sustainability product. The current AI-power cycle has created a much larger opportunity because customers increasingly value speed, reliability and power availability above a narrow comparison of fuel cost alone.

The company should not be confused with a regulated utility. Bloom sells equipment, installation, service and electricity arrangements. Its revenue can be project-driven and lumpy. Large quarters may depend on system acceptances and customer schedules. That creates more volatility than a recurring software company despite the long-duration service relationships.

3. Solid Oxide Technology: What Bloom Actually Sells

Bloom’s Energy Server is a modular solid oxide fuel cell system. A fuel cell uses an electrochemical reaction to convert fuel into electricity. It does not burn fuel in the same way as a turbine or reciprocating engine, although the upstream production and use of natural gas still create greenhouse-gas emissions.

Solid oxide cells operate at high temperatures. That allows Bloom’s system to use multiple fuels, including natural gas, biogas and hydrogen. The same core platform can also operate in reverse as a solid oxide electrolyzer, using electricity and heat to split water into hydrogen.

Why customers may prefer onsite solid oxide power

• Speed: modular systems can be deployed faster than large grid upgrades or conventional generation projects.
• Reliability: onsite generation can reduce exposure to grid outages and transmission constraints.
• Power density: Bloom has cited deployments capable of approximately 100 MW per acre.
• Modularity: customers can add capacity in blocks rather than wait for a single large plant.
• Fuel flexibility: systems can use natural gas today and potentially lower-carbon fuels in the future.
• Water profile: fuel cells generally use less operational water than thermal generation and many cooling-intensive alternatives.

The central environmental tradeoff

Bloom often describes its natural-gas Energy Servers as cleaner than conventional grid power or combustion-based alternatives. That can be true in specific locations and operating conditions, especially when avoided transmission losses, local pollutants and grid mix are considered. The systems are not automatically zero-carbon when operating on natural gas. Investors should separate lower-carbon power, renewable biogas, carbon capture and fully hydrogen-powered operation.

This distinction matters for data-center permitting. Communities may accept onsite power because it reduces grid strain, but they may also challenge local emissions, gas infrastructure, noise, land use and the effect of rapid AI growth on electricity prices.

4. Business Model and Revenue Mix

Bloom reports revenue in four categories: product, installation, service and electricity. Product sales are the largest and highest-impact category. Installation can be low-margin or negative-margin because project complexity, labor and site conditions vary. Service provides longer-duration revenue and is strategically important because every installed system requires maintenance and replacement activity. Electricity revenue arises from arrangements where Bloom or a financing structure retains ownership and sells power.

Q2 showed why consolidated margin can be misleading. Product gross margin was strong, while installation remained negative. Product costs also benefited from a $37.4 million recovery of previously paid import tariffs. Excluding that recovery mechanically—not as a company-reported adjusted metric—consolidated gross margin would have been roughly 29.9%, so the 33.4% headline should not automatically be treated as a clean run-rate.

Bloom’s business also depends on acceptance milestones. Systems may be manufactured and installed before all revenue can be recognized. Contract assets, inventories and customer deposits can therefore rise quickly during expansion. Working-capital discipline is an important part of the thesis.

5. The AI Data-Center Power Thesis

Power availability has become a limiting factor for AI infrastructure. Hyperscalers and data-center developers can acquire land, servers and capital but still wait years for utility interconnection, generation and transmission upgrades. Bloom’s value proposition is to bring power directly to the site.

The company’s 2026 surveys found that developers increasingly expect to use onsite generation, with 61% of surveyed decision-makers saying they planned to bring their own power if the grid could not meet requirements. Company-sponsored surveys are not independent market forecasts, but they describe a constraint visible across utility filings, data-center announcements and regional development debates.

Bloom can benefit even if the long-term grid eventually catches up. Onsite systems may operate as bridge power, permanent primary generation, microgrids or supplemental capacity. Customers care about the economic value of bringing compute online earlier. A data center that begins operating months or years sooner may justify a higher power cost than a conventional industrial user would accept.

Why the opportunity could be structural

• AI workloads are increasing the power density of individual campuses.
• Utilities face generation, transmission and interconnection bottlenecks.
• Data-center revenue can begin only after power becomes available.
• Onsite systems can reduce dependence on one grid connection.
• Bloom can deploy modular capacity in stages as compute demand grows.

Why the cycle can still disappoint

AI capital expenditure can change quickly. Customers may redesign projects, delay campuses, secure utility power or choose turbines, engines, batteries, nuclear contracts or other onsite technologies. Large announced frameworks can take years to convert and may depend on customer financing, site approval and fuel availability.

6. Oracle: Up to 2.8 GW, but Read the Agreement Carefully

In April 2026 Bloom announced an expanded strategic partnership with Oracle. The master agreement supports procurement of up to 2.8 GW of Bloom fuel-cell systems, with an initial 1.2 GW described as deploying across Oracle projects in the United States. The agreement follows a prior installation that Bloom said became operational in 55 days.

The scale is strategically important. A 2.8 GW framework is larger than Bloom’s current annual production capacity and supports the decision to expand Fremont. It also validates Bloom’s relevance to a major cloud and AI infrastructure operator.

Investors should not treat the entire 2.8 GW as immediately recognized backlog or guaranteed near-term revenue. Master agreements define commercial terms and potential volume. Actual revenue depends on project orders, site selection, delivery schedules, acceptance and customer decisions.

The Oracle warrant

Bloom issued Oracle a warrant for up to 3,531,073 shares at $113.28 per share on April 9, 2026. Oracle completed a cashless exercise on May 1. Bloom issued 1,905,433 shares on a net basis and another 248,798 inducement shares because net settlement reduced the share issuance relative to gross settlement.

The aggregate fair value of the issued shares, including the inducement, was $324.4 million. Bloom accounts for the arrangement as consideration payable to a customer’s customer and recognizes it as a reduction of revenue as underlying systems are delivered. The exercise is complete; October 9 is no longer a live warrant-expiration catalyst.

7. Brookfield: Financing Platform, Not a $25 Billion Purchase Order

Bloom and Brookfield announced a $5 billion AI infrastructure financing framework in October 2025. On June 30, 2026 they expanded the framework to $25 billion. Brookfield can finance qualifying onsite-power projects, reducing the need for Bloom or the customer to fund the entire system directly.

This is strategically valuable because financing has historically been a bottleneck for distributed generation. A sophisticated infrastructure partner can own project assets, provide capital and allow Bloom to sell equipment into financed structures.

Bloom and Brookfield use joint-venture structures. Bloom sells systems to the ventures and recognizes equity-method accounting effects, including deferral of intra-entity profit over the assets’ depreciable lives. This can make GAAP financial statements more complex than a simple equipment sale.

The Q2 filing also disclosed a project-option arrangement linking Oracle and a Brookfield vehicle. Bloom paid $50 million in May for option-related rights and assigned them to the Brookfield vehicle in July. If Brookfield proceeds with the underlying project, it is expected to pay $50 million to Bloom; if not, the contractual structure is designed to return the rights and recover Bloom’s payment. This is a financing and project-rights structure—not a new $50 million revenue event or a new purchase order.

The quality of the partnership should be evaluated through funded projects, megawatts deployed, customer names, cash conversion and returns—not only the headline framework size.

8. AEP, Equinix, CoreWeave and the Broader Customer Base

Bloom’s data-center credentials do not rest on Oracle alone. The company signed a procurement agreement with American Electric Power for up to 1 GW, beginning with a 100 MW order. AEP can use Bloom systems to supply data-center customers in areas where traditional infrastructure cannot arrive quickly enough.

Bloom’s relationship with Equinix surpassed 100 MW in 2025, with approximately 75 MW operating and another 30 MW under construction at that time. The companies have worked together for roughly a decade across multiple U.S. data centers. This history supports the claim that Bloom technology can operate in mission-critical environments.

CoreWeave selected Bloom to provide onsite power for a high-performance computing data center in Illinois. Bloom also serves semiconductor, retail, healthcare and industrial customers.

Customer concentration remains material

The amended Q2 Form 10-Q is unusually important. Bloom corrected an inversion in the original filing: during Q2, one non-related customer represented approximately 73% of total revenue. Across the first half, two customers represented approximately 44% and 21% of revenue. At June 30, three customers represented 36%, 34% and 17% of accounts receivable—87% in aggregate—although Bloom reported no material credit losses from them.

The word “customer” can refer to a contractual project-financing counterparty rather than the ultimate data-center operator. Investors should not automatically label the 73% customer as Oracle, Brookfield, AEP or another named end user without an explicit filing. The 10-Q attributes product growth to a significant deployment for a large AI infrastructure customer and multiple projects executed through the Brookfield joint venture.

Related-party revenue was only $2.8 million in Q2 but $376.1 million for the first half, reflecting the concentration of Brookfield-JV deliveries in Q1. This shift does not remove concentration risk; it shows how sharply the contractual channel can change between quarters.

A small number of hyperscale customers can accelerate growth and improve factory utilization. They can also gain negotiating leverage, delay projects or create volatility when one acceptance milestone moves between quarters.

MiTAC: what the 250 MW number is made of

On August 6, 2026 Bloom announced an expanded partnership with MiTAC Computing Technology Corp. Bloom will deploy fuel cell systems for an islanded microgrid at MiTAC’s AI server manufacturing campus in Fremont, California, supporting current operations and future AI server production, and building on an existing Bloom installation at MiTAC’s San Jose facility. The disclosure travelling with it matters more than the single site: Bloom stated it now serves nearly two dozen AI infrastructure customers representing roughly 250 MW of capacity, against almost none two years ago.

Two readings sit on top of each other. The 250 MW is an aggregate across the whole AI customer base rather than a MiTAC figure, and it describes capacity served rather than contracted backlog. MiTAC is also a different kind of customer from Oracle or Brookfield: it builds the servers rather than operating the data centre, which extends the onsite-power case one step up the supply chain, into the factories that make the hardware.

9. Financial Performance: From Promise to Operating Leverage

Bloom reported full-year 2025 revenue of $2.02 billion, up 37.3% from 2024. GAAP gross margin improved to 29.0%, non-GAAP operating income reached $221.0 million and operating cash flow was positive for a second consecutive year.

Q2 2026 accelerated again. Revenue rose 165.5% year over year to $1.065 billion, driven by a 215.4% increase in product revenue. GAAP operating income reached $182.2 million versus a $3.5 million loss a year earlier. Operating cash flow was $226.4 million, bringing first-half operating cash flow to $300.0 million.

The company is producing real operating income, but investors should not rely only on headline or non-GAAP metrics. Q2 stock-based compensation was $56.4 million. The quarter also benefited from the tariff recovery, while customer deposits supported working capital and equity-method accounting tied to joint ventures complicates comparison between product sales, cash conversion and net income.

10. Raised 2026 Guidance and the Execution Test

After Q2, Bloom raised 2026 revenue guidance from $3.4–$3.8 billion to $3.9–$4.2 billion. The new $4.05 billion midpoint is 12.5% above the prior midpoint and implies approximately 100% growth over 2025. Non-GAAP gross-margin guidance remained about 34%, while the operating-income and EPS outlooks moved sharply higher.

What matters after the headline beat

• Product megawatts accepted and recognized.
• Contribution from Oracle, AEP and Brookfield-funded projects.
• Product versus installation mix.
• Service margin and warranty accruals.
• Customer deposits and contract assets.
• Inventory growth relative to future deliveries.
• Progress on the 2 GW factory expansion.
• Any quantified update to company-defined backlog and firm orders.

11. Backlog, Performance Obligations and Revenue Visibility

Bloom’s latest absolute backlog disclosure remains the year-end 2025 figure: approximately $20 billion of total current backlog, including roughly $6 billion of product backlog. Management said on the Q2 call that backlog was growing faster than revenue, but it did not publish a new absolute Q2 value. The year-end figure must therefore not be presented as a June 30 balance.

Company-defined backlog is broader than GAAP remaining performance obligations. At June 30, 2026, Bloom reported $442.4 million of unsatisfied product and installation performance obligations expected to be recognized within one to two years, plus $51.7 million mainly related to service contracts recognized over one to 25 years. These GAAP figures are not a substitute for the company’s broader backlog definition.

Customer deposits increased from $78.2 million at year-end 2025 to $360.6 million at June 30, while total deferred revenue and customer deposits reached $445.0 million. The filing attributes the $282.4 million deposit increase mainly to recently executed agreements and milestone payments. That is constructive evidence of customer commitment, but contract assets also climbed to $428.3 million and accounts receivable to $458.1 million, so investors should watch both sides of the working-capital bridge.

July 2026 short-seller dispute

On July 8, Hunterbrook Media published a report questioning Bloom’s accounting, backlog presentation and access to scandium oxide. Hunterbrook disclosed that affiliated investment activity could benefit from a decline in Bloom’s share price. On July 9, Bloom filed an 8-K categorically rejecting the report’s accounting claims as false and misleading, stating that it stood behind its audited financial statements and SEC reporting.

Bloom also said it had sufficient scandium oxide for current demand and backlog, was not dependent on China for that supply, and had visibility to support substantially greater future production. The company response is an official disclosure, but it does not represent an independent adjudication of every issue raised. Investors should compare the company-defined backlog with GAAP performance obligations, monitor working-capital conversion and review any future auditor, regulator or customer disclosures.

Hunterbrook published a follow-up on July 30 focused on efficiency, output and longevity claims. It remains an interested, short-biased source rather than an adjudicated finding. Bloom had not filed a new formal response specifically addressing that follow-up by the August 2 evidence cut-off. Shareholder-law-firm notices and litigation allegations also began circulating; they should be treated as claims, not proof of wrongdoing, unless a court, regulator or audited filing establishes otherwise.

12. Manufacturing Expansion and Supply-Chain Execution

Bloom’s prior public target is to expand the Fremont manufacturing facility from an annual production run rate of 1 GW to 2 GW by the end of 2026. Q2 management did not publish a new numeric capacity target. It said the company was not currently gated by capacity or supply chain and described a “Copy Exactly” approach to adding production capability. That is a qualitative execution claim, not a replacement for the 2 GW target or proof of installed capacity.

The same facility could potentially support up to approximately 5 GW of annual capacity. Bloom estimated that each additional 1 GW increment could require six to nine months and approximately $100–$150 million of capital expenditure.

Why factory execution matters

Demand can exceed revenue if Bloom cannot secure stacks, power electronics, structural components, field labor and installation capacity. Capacity expansion must preserve product yield, reliability and margin. A rushed scale-up can produce warranty costs or service problems that appear after deployment.

Inventory rose to $758.2 million at June 30 from $643.3 million at year-end. First-half purchases of property, plant and equipment reached $77.8 million, versus $21.5 million a year earlier. The inventory build may support planned deliveries and protect lead times, but it also ties up cash and creates risk if customer schedules change or product designs evolve.

Supply-chain exposure

Bloom said its direct supply chain was not dependent on China, but tier-two and tier-three suppliers use rare-earth materials and electronic components sourced from China. Tariffs, geopolitical conflict, freight costs and commodity prices can affect margins. The company uses supplier diversification, strategic inventory and contractual pass-through mechanisms, but those tools may not fully offset rapid cost increases.

13. Service Economics, Reliability and the Installed Base

Service has historically been a weak point for many fuel-cell companies because field maintenance and stack replacement can consume the margin earned on the initial sale. Bloom’s Q2 service gross margin improved to 18.7% from 9.2% a year earlier, while non-GAAP service margin reached 22.0%.

This improvement is strategically important. A growing installed base should produce recurring service revenue, but only if maintenance costs, failure rates and replacement cycles remain controlled. Customers buying mission-critical data-center power require high availability. A system failure can have consequences far beyond the replacement cost.

Investors should monitor accrued warranty, service gross margin, replacement expense and customer renewals. Accrued warranty rose to $77.8 million at June 30 from $20.0 million at year-end, while Q2 performance-guarantee costs increased by $3.0 million because of fleet degradation. Rising accruals are not automatically negative, but they should remain proportionate to deployments and fleet age.

Bloom also disclosed performance guarantees under operation-and-maintenance agreements. The aggregate caps were approximately $846.1 million, with approximately $468.9 million of remaining potential payments at June 30, 2026. These figures are not a current recognized loss or a prediction that the amounts will be paid; they define contractual exposure if systems underperform specified output or efficiency guarantees.

14. Hydrogen and Electrolyzer Optionality

Bloom’s solid oxide platform can operate in reverse as an electrolyzer. High-temperature electrolysis can use both electricity and heat, reducing electricity consumption compared with lower-temperature systems in suitable industrial settings. Bloom has demonstrated a 4 MW electrolyzer at NASA Ames and built dedicated capacity in Delaware.

The hydrogen business remains optionality rather than the main current valuation driver. Bloom’s own annual report states that hydrogen infrastructure and supply have not yet developed enough to have a significant impact on the market. Pipelines, transport, storage and clean-power economics remain constraints.

The electrolyzer may become more valuable in industries with waste heat, nuclear steam or abundant renewable power. Bloom also offers hydrogen-powered Energy Servers. Those capabilities create a potential closed loop: produce hydrogen when clean electricity is available, store it and convert it back to electricity later. The economics remain project-specific.

15. Korea and International Business

South Korea has been an important market through Bloom’s relationship with SK ecoplant. The companies operate a Korean joint venture and maintain distribution agreements that include purchase commitments through 2027. SK can satisfy some commitments using Energy Servers or electrolyzers and pursue projects outside Korea.

The Korean venture adds local manufacturing and market access. It also creates accounting complexity and historical concentration. SK reduced its ownership stake in Bloom during 2025 and was no longer treated as a related party after July 10, 2025.

International growth can diversify Bloom beyond U.S. data centers, but local policy, fuel prices, carbon regulation, currency and project finance influence demand. European data-center opportunities may be attractive because grids are constrained, but natural-gas infrastructure and carbon rules can be less favorable than in some U.S. regions.

16. Cash, Debt, Convertibles and Dilution

Bloom reported $2.667 billion of unrestricted cash and cash equivalents at June 30, 2026. That large balance resulted partly from a $2.5 billion zero-coupon convertible-note offering completed in late 2025. Total unpaid debt principal was $2.530 billion and the balance-sheet carrying value was $2.478 billion, leaving modest net cash against principal before lease and financing obligations.

Bloom has liquidity to fund capacity, inventory and working capital, but the cash balance should be read together with convertibles, financing obligations and a rapidly growing asset base. Convertible instruments reduce cash interest but create dilution when the share price rises.

Bloom reported 294,527,346 common shares outstanding as of July 22, 2026. Q2 weighted-average diluted shares were 323.3 million, reflecting convertibles, options, awards and the warrant effect. That diluted denominator is more relevant than the basic share count when evaluating per-share economics at elevated stock prices.

Stock-based compensation

Q2 stock-based compensation was $56.4 million, up 86.9% year over year. At June 30, Bloom also had more than 10.1 million RSUs and PSUs outstanding under the 2018 plan, in addition to options and other equity programs. These expenses are non-cash in the quarter, but the underlying shares transfer economic value and dilute owners.

17. Valuation Context: A Great Business Can Still Be an Expensive Stock

Using the July 31 close of $205.81 and 294.53 million shares outstanding on July 22, Bloom’s simple equity value is approximately $60.6 billion. Using the Q2 diluted share denominator of 323.3 million produces approximately $66.5 billion—consistent with the market-data capitalization snapshot. Adding $2.530 billion of debt principal and subtracting $2.667 billion of cash leaves enterprise value close to equity value before lease, financing-obligation and other adjustments.

Against the $4.05 billion midpoint of updated 2026 revenue guidance, this is roughly 14.9 times enterprise value to sales on the basic share count and about 16.4 times on the Q2 diluted denominator. These are simple spot calculations, not target prices, but they show how much multi-year execution remains embedded in the stock.

The valuation can be supported only if Bloom grows into it. A simple peer multiple is difficult because Bloom combines electrical equipment, distributed generation, data-center infrastructure and clean-energy exposure. Investors should model revenue, product gross profit, service economics, capital requirements and diluted shares rather than focus on the nominal stock price.

Post-Q2 technical map

BE closed at $205.81 on July 31. The stock stood above its 5-day and 200-day moving averages but below its 20-day and 50-day averages, while 14-day ATR was near $30. That combination describes a violent rebound inside a damaged intermediate trend—not a stable momentum setup.

Technical levels are observations, not forecasts. Moving averages and volatility measures change daily; use the live Finviz chart above before publication or trading decisions.

Analyst reaction: strong quarter, divided valuation views

Post-Q2 research was not uniformly bullish. Mizuho upgraded BE to Outperform while lowering its target to $242, citing better execution, margin expansion and the pullback. Bernstein raised its target to $282 with a Market Perform view. Other firms maintained positive ratings but cut targets, while Jefferies and Wells Fargo stayed more cautious. The dispersion reinforces the central debate: operating momentum improved faster than confidence in concentration-adjusted valuation.

18. Competitive Landscape

Bloom competes with the grid, gas turbines, reciprocating engines, battery systems, other fuel cells, utility procurement and emerging nuclear solutions. Its competitive advantage is the combination of speed, modularity, reliability and power density.

Plug Power is not the closest financial peer. PLUG is more exposed to PEM fuel cells, hydrogen production and material handling. Bloom’s current economics are dominated by stationary onsite power and data-center infrastructure.

19. Management, Governance and Execution

KR Sridhar is Founder, Chairman and CEO. Founder leadership provides technical continuity and a long-term strategic vision. It can also concentrate influence. Investors should monitor board oversight, capital allocation and disclosure discipline during rapid expansion.

Simon Edwards became Chief Financial Officer in April 2026. He previously served as CFO and later CEO of Groq and held finance roles at technology companies and GE. His appointment is strategically relevant because Bloom is shifting from product validation to gigawatt-scale capital and infrastructure execution.

The leadership team includes Chief Commercial Officer Aman Joshi, Chief Technology Officer Ravi Prasher, Chief Operations Officer Satish Chitoori and executives responsible for quality, service, legal, policy and manufacturing.

In June 2026 Bloom filed an amended 8-K describing a new performance-based equity award to Dr. Sridhar with 319,082 target shares. The award is tied primarily to aggregate revenue over a multi-year performance period and includes holding requirements. It can align management with sustained growth, but it also adds potential dilution and makes the quality—not merely the amount—of future revenue an important governance consideration.

Management scorecard

• Convert large frameworks into firm, profitable orders.
• Complete the 2 GW expansion on time and within budget.
• Preserve product and service reliability during rapid scaling.
• Reduce installation losses.
• Control stock-based compensation and diluted share growth.
• Communicate backlog definitions and customer concentration transparently.
• Balance natural-gas growth with credible decarbonization pathways.

20. Bloom Energy Catalyst Calendar

21. Bull, Base and Bear Scenarios

22. Risk Register and Monitoring Checklist

Post-Q2 monitoring checklist

• Refresh price, basic shares and diluted share assumptions before using valuation multiples.
• Separate confirmed product backlog from framework announcements.
• Track product, installation, service and electricity margins separately.
• Normalize Q2 margin for the $37.4M tariff recovery before assuming a run-rate.
• Compare $758.2M inventory and $428.3M contract assets with deposits and operating cash flow.
• Seek numeric evidence that the prior 2 GW capacity target remains on schedule.
• Ask how much Oracle, Brookfield and other large-project volume is included in updated guidance.
• Track whether the 73% Q2 customer concentration falls as new projects enter revenue.
• Compare any backlog update with GAAP performance obligations and the issues raised in the July short-seller dispute.
• Monitor convertible-note dilution, SBC, options and RSU/PSU issuance.

23. How to Read the Next Earnings Release

Bloom’s Q2 earnings looked spectacular at the headline level, but project acceptance can cause large revenue changes. The next review should begin with repeatability and the operating bridge.

A revenue beat accompanied by lower deposits, rising inventory and delayed customer schedules may be lower quality than a smaller beat supported by cash, firm orders and margin. Management’s explanation of framework conversion will be central because the stock’s valuation is based on future scale.