Bloom Energy AI Data-Center Deals, November 2026: 900 MW NTT DATA Agreement Headlines a Multi-Gigawatt Pipeline
Bloom Energy closed at $195.73 on November 6, 2026, up 13% with no specific catalyst identified in published reports, as a 900 MW NTT DATA fuel-cell agreement and a 702 MW Google-backed Wyoming project anchored a disclosed pipeline now spanning several gigawatts.
- 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 shares closed at $195.73 on November 6, 2026, a gain of 13% on that Friday. No published report has attributed the move to a specific event on that date. What is clear is that the company entered November with a rapidly expanding commercial profile: coverage from that month includes the announcement of a 900 MW fuel-cell supply agreement with NTT DATA for onsite power generation across multiple data-center projects — widely reported as Bloom's largest overseas commitment to date — and a Barclays analyst action that upgraded the stock to Overweight while simultaneously cutting the price target. That combination of enlarged commercial reach and mixed analyst conviction captures the central tension in Bloom Energy's current story.
The NTT DATA agreement is the latest addition to a pipeline of AI-infrastructure power deals that has accumulated at an unusually rapid pace. First deliveries under the NTT DATA contract are reportedly expected next year. In November 2026, Google announced a $17 billion Wyoming data-center complex with initial capacity of 2.7 GW, with Bloom Energy cited as the supplier of 702 MW of natural-gas fuel cells incorporating carbon capture; Google had separately signed a 20-year power purchase agreement with AEP covering that same 702 MW project, with AEP reportedly expected to invest approximately $4 billion. Oracle, meanwhile, reaffirmed a 2.4 GW fuel-cell commitment in September 2026 after a force-majeure notice was resolved — the CEO described Bloom's energy servers as fungible within Oracle's data-center power architecture. California Resources and Bloom Energy are reported to be exploring a 1 GW fuel-cell deployment in California, linking carbon capture and sequestration to the generation assets. Nebius Group powered up its first Bloom Energy fuel-cell block at a 300 MW New Jersey data center in October 2026. The pattern — hyperscalers, cloud operators, and infrastructure investors signing multiyear agreements at gigawatt scale — represents a structural shift from the distributed-commercial model that defined Bloom's first two decades.
Bloom Energy was founded in 2001 and listed on the NYSE in 2018, spending most of its public life selling solid-oxide fuel cells to hospitals, universities, and corporate campuses. Its inclusion in the S&P 500, reported in September 2026, reflected how dramatically that profile has changed. The company reported record third-quarter 2026 financial results in October and raised its full-year 2026 guidance, with an 8-K filed on October 22, 2026 announcing those results. At the fiscal 2025 level, Slicast's XBRL-derived capex data shows capital expenditure of $57 million, down 3.6% from $59 million in fiscal 2024, representing just 3% of $2.00 billion in revenue. Within Slicast's eight-company power-sector peer group, that places Bloom seventh of eight by capex intensity and sixth of eight in absolute capex terms — a lean manufacturing-and-deployment profile that allows the pipeline to grow without immediate proportional investment in fixed assets, but that also makes supply-chain and manufacturing throughput the binding constraint as order volumes scale. LS Electric's $139 million Bloom Energy supply order for a U.S. data center, reported in October 2026, is one visible sign of that supply chain beginning to mobilize.
The optimism surrounding these commitments coexists with material friction. Within November 2026, coverage also recorded a session in which Bloom Energy fell 14%; separate reporting attributed the decline to investor concerns about data-center construction delays, and MarketWatch listed the company among the S&P 500's largest single-session losers amid those same fears. Prior to that, in September 2026, coverage noted a key data-center deal stalling; the Oracle force-majeure episode, though resolved, demonstrated that hyperscaler commitments can be subject to renegotiation; and legal uncertainties were noted without specific details publicly disclosed in available reporting. Barclays' decision in November to upgrade the stock while cutting its price target conveys a nuanced signal: the directional thesis has improved, but the valuation debate remains open. The CEO's framing of NIMBY opposition to conventional data centers as a commercial opportunity for Bloom's onsite model is strategically coherent, but it simultaneously acknowledges that planning and community-acceptance headwinds are real, sector-wide constraints.
Three concrete signals will determine whether November's 13% session advance proves durable or speculative. First, the delivery timeline for the NTT DATA 900 MW agreement: first deliveries are reportedly expected in 2027, and any disclosed schedule slippage would likely reprice execution risk across the broader pipeline. Second, AEP's progress on the Wyoming 702 MW project: with a 20-year Google PPA and roughly $4 billion in reported investment commitments, the structural backing is unusually strong, but regulatory and financing milestones will serve as a leading indicator of whether large-scale Bloom deployments can proceed at the pace the pipeline implies. Third, Oracle's Project Jupiter trajectory: at 2.4 GW it remains the single largest disclosed commitment, and any renewed contractual or legal disruption would carry outsized signal value for the entire data-center fuel-cell investment thesis. Underlying all three is a quieter financial question: whether Bloom can sustain its record Q3 2026 momentum without a capex surge that breaks from its historically lean model — currently running at 3% of revenue — as gigawatt-scale manufacturing demands begin to materialize. The demand signal is unambiguous; the execution proof is still ahead.