Bloom Energy Stock Decline, November 2026: 15% Single-Session Drop Amid AI Power Sector Sell-off
Bloom Energy closed at $142.39 on November 20, 2026, down 15.3% on the session and roughly 26% over the week according to TIKR.com, as a broad AI-infrastructure sector sell-off hit the stock despite a 900 MW NTT DATA agreement and record Q3 2026 results filed with the SEC.
- 최근 회계연도 설비투자
- $57M (FY2025)
- 전년 대비
- -3.6% · $59M → $57M
- 투자 / 매출
- 3% (FY2025, $2.00B)
- 최고 기록 기간
- $27M · 2023-03-31
Bloom Energy's shares fell 15.3% on November 20, 2026, closing at $142.39, and ranked among the S&P 500's biggest decliners on the session, according to MarketWatch. TIKR.com reported that the stock had shed roughly 26% over the week, part of a broader rout across AI power equities: The Motley Fool noted GE Vernova and Vertiv each lost between 10% and 24% over the same stretch, while FuelCell Energy and Plug Power fell sharply in the period as well, per 24/7 Wall St. TIKR.com attributed the broader sector pressure to turbulence around Nvidia; MarketWatch cited investor concerns about data-center construction delays; and Simply Wall St flagged a short report challenging Bloom's AI data-center narrative as an additional overhang. While each of those factors may have contributed to the week's losses, no source in Slicast's coverage provides a definitive account of what drove the specific magnitude of November 20's session decline.
The selloff arrives against one of the more commercially active stretches in Bloom's history. The company filed a 10-Q on October 22, 2026 and an 8-K on the same date announcing record third-quarter 2026 results, simultaneously raising its full-year 2026 guidance. FY2025 revenue reached $2.00 billion, against which Bloom recorded $57 million in capital expenditure — a 3% intensity ratio, down marginally from $59 million in FY2024, a year-over-year decline of 3.6%. Among the eight power-sector companies Slicast tracks, Bloom ranks seventh by capex intensity, behind category leader Vistra, and sixth by absolute capex spend, behind category leader Constellation Energy. The lean capex profile is a structural feature of Bloom's model: the company manufactures solid-oxide fuel-cell stacks and delivers them as configured onsite power units, rather than building generation infrastructure in the conventional utility mold. That posture limits balance-sheet capital consumption while concentrating execution risk in production throughput and the speed at which customer sites can be commissioned.
The commercial pipeline that animates the bull case is unusually visible. In November 2026, NTT DATA and Bloom announced a 900 MW agreement covering onsite power generation across multiple data-center projects — described by Yahoo Finance Australia as the company's largest overseas win to date, with initial deliveries reportedly expected in the following year. The Wyoming thread adds further scale: in October 2026, Google and AEP were reported to have signed a 20-year power purchase agreement for a 702 MW fuel-cell project with Bloom as equipment supplier, with AEP reported to be investing approximately $4 billion in the build; a separate report described Google planning a $17 billion Wyoming data-center campus with an initial draw of 2.7 GW. California Resources and Bloom have reportedly been exploring 1 GW of fuel-cell deployment across California, to be paired with carbon capture and sequestration. At the operational end of the pipeline, Nebius Group energized the first Bloom fuel-cell block at its 300 MW New Jersey data center as of reported October 2026. LS Electric, in October 2026, was reported to have won a $139 million Bloom-related order for a U.S. data center, illustrating supply-chain activation further downstream.
An earlier episode demonstrates how quickly execution concerns can reshape the narrative. Reports from September 2026 indicated that a key data-center deal had stalled and that force-majeure language in Bloom's Oracle arrangement had unsettled investors; September 2026 reporting also indicated that the force-majeure issue had resolved, clearing the path for what reporting described as a 2.4 GW Oracle buildout. Oracle subsequently proposed up to $55 million in annual power-bill credits, and Bloom's CEO publicly characterized the company's energy units as "fungible" within Oracle's data-center power architecture. The episode illustrates a structural feature of Bloom's risk profile: multi-year, multi-hundred-megawatt contracts provide substantial revenue visibility but also concentrate exposure — if a major customer revises timelines or invokes contractual protections, the effect on reported backlog can be outsized. Barclays navigated that tension in November 2026 by upgrading Bloom to Overweight while simultaneously cutting its price target, a split signal reflecting genuine disagreement on valuation after the stock's run. RBC is separately reported to see Bloom as a potential equipment provider for large Panama AI data-center projects, while Simply Wall St has flagged a short report challenging Bloom's AI data-center narrative.
Bloom's CEO argued publicly in November 2026 that neighborhood and regulatory opposition to large utility-scale generation and high-voltage interconnection projects is a structural tailwind for distributed fuel cells, which can be sited onsite without entering long interconnection queues. That argument gains texture from the breadth of the announced pipeline — spanning hyperscalers, enterprise data-center operators, and co-located utility-scale projects — and from TIME magazine's recognition of Bloom among America's top manufacturers for 2027. Three concrete signals merit monitoring as the investment case develops: whether full-year 2026 results, when disclosed, confirm or revise the raised guidance; whether the NTT DATA 900 MW deployment commences on the reported timeline, since that contract's execution will test the production throughput the bull case requires; and how the Google-AEP Wyoming and Oracle projects progress from signed agreements to energized capacity — the gap between contract announcement and commissioning has historically been where fuel-cell project schedules slip. The 15.3% session loss on November 20 is a reminder that in the current AI-infrastructure cycle, market repricing can run materially ahead of changes in underlying fundamentals, in either direction.