BE · Bloom Energy Corporation

Current Thesis

On-site-power-for-AI is the dominant leg: Bloom's solid-oxide fuel cells (SOFC) are the fastest behind-the-meter fix for the data-center grid-interconnect bottleneck, and two marquee contracts in six weeks turned the story from "perpetually-unprofitable green-energy hope" into an AI-infrastructure supplier with a ~$20B backlog. The Oracle "Project Jupiter" agreement (2026-04-27, up to 2.8 GW) and the Nebius 10-year/$2.6B master deal (2026-05-20, 328 MW first phase) are the proof points, layered on a Q1 turn to GAAP profit. The theme (industrial-power-AI) is ACCELERATING and cluster-confirmed. The stock itself, however, is in a MATURING/digesting leg: ~$266 on 2026-06-05 versus a $322.83 ATH set 2026-05-20, now back at the 27-analyst consensus PT ($263.13). This is momentum-to-be-bought-on-a-retest, not a fresh breakout to chase.

Bull Case

• Oracle Project Jupiter (2026-04-27): Oracle + BorderPlex Digital Assets committed to up to 2.8 GW of Bloom SOFC (1.2 GW already in progress) to fully power an AI campus in Doña Ana County, NM (up to 2.45 GW installed at the site). BE jumped +25.7% on the news. This is the defining transaction it locks up a large share of Bloom's production capacity and validates SOFC for mission-critical AI workloads.
• Nebius master agreement (2026-05-20): 10-year, up to $2.6B hardware+service deal; 328 MW first phase (~250 MW guaranteed), first project live in 2026. Daiwa moved Hold→Outperform, $324 PT, same day.
• Backlog ~$20B (per Q1 2026): multi-year revenue visibility, no longer a single-deal story.
• Q1 2026 print (2026-04-28) was a regime change: revenue $751.1M, +130.4% YoY; first real GAAP net income ($70.65M); adj EPS $0.44; non-GAAP gross margin 31.5%; adj EBITDA $143M; operating cash flow +$73.6M.
• Guidance raised: FY2026 revenue $3.40–$3.80B; midpoint growth guide lifted to ~80% YoY from a prior ~60%; non-GAAP operating income $600–$750M; non-GAAP EPS $1.85–$2.25. CEO has framed the build-out as self-funding ("the AI boom is funding itself").
• Cluster intact (2026-05/06): VST, CEG, GEV, OKLO, IREN and Nebius all ripping on the AI-power-crunch story; recurring "industrials whale alerts" repeatedly flag BE through early June institutional flow building.
• RBC reiterated Buy, $335 PT the high mark still sits above the current price.

Bear Case

• Price is back AT consensus, not below it: ~$266 (2026-06-05) versus the 20-to-27-analyst average PT of $263.13. The easy re-rate is done; average implies roughly flat-to-modest upside, and the low target of $55 shows how wide the dispersion is on a narrative multiple.
• Rolling off the high: $322.83 ATH (2026-05-20/21) → $291.37 prior close → ~$266 (2026-06-05), about -17.5% off the ATH. The vertical leg is consolidating, momentum decelerating.
• Insider distribution: director Mary Bush filed to sell 25,000 shares into the strength a sell-the-rip signal from inside.
• Valuation is pure narrative: forward P/E ~120x; GuruFocus GF Value pegs fair value near $26 versus a ~$266 price. Any deal-flow pause or margin wobble de-rates violently at this multiple.
• Concentration/lumpiness: the backlog is dominated by a handful of mega-deals (Oracle, Nebius). One slipped install timeline, or a hyperscaler choosing gas turbines (GE Vernova) over SOFC, reframes the "Bloom is the answer" thesis.
• Beta ~3.75: expect 2-4x SPY daily moves in both directions; this amplifies a capex-unwind drawdown.

Setup & Price Structure

• Last (2026-06-05): ~$266.26; prior close $291.37 a roughly -8.6% leg down into June.
• ATH: $322.83 (2026-05-20/21, Nebius spike) → currently ~-17.5% off the high, in a post-spike digestion that has now given back the $287–$291 area.
• Range/return: 52-week $18.12 → $322.83; +214% YTD; ~+1,380% over one year; market cap ~$80B+.
• Key shelf: the pre-Nebius breakout base sits $230–$250. A higher-low retest into that zone is the clean momentum entry; chasing into the roll-off from $322 is not.
• Read: MATURING leg inside an ACCELERATING theme. The question the tape is answering now is whether $250–$260 holds as support or the digestion deepens toward the $230–$240 shelf.
• Beta 3.75 size for amplified swings.

Catalyst Calendar (next 30 days)

• No confirmed dated catalyst inside the 2026-06-06 → ~2026-07-06 window. Trade the structure, not an event.
• Q2 2026 earnings est. ~2026-07-30 (OUTSIDE the 30-day window): the next binary. Flag a blackout window ~3 trading days prior given the ~120x forward multiple.
• Oracle Project Jupiter / first Nebius 328 MW project "operational in 2026," no fixed date: watch for ground-breaking or first-power milestone PRs as theme-confirming catalysts.
• Director Form 4 (Mary Bush, 25,000 sh): watch for the executed filing as a distribution datapoint.
• Ongoing flow: recurring "industrials whale alerts" naming BE monitor for flow accelerating vs. fading.

What Would Change Our Mind

• Bull re-acceleration (upgrade trigger): reclaim $291–$300 on expanding volume, or a new hyperscaler SOFC contract on the scale of Oracle/Nebius confirms the theme is still in its early innings and a fresh-breakout entry is back on.
• Thesis-break / invalidation: a weekly close below the rising 20-EMA / the $225–$240 pre-breakout shelf; OR an FY2026 revenue guide cut below the $3.40B floor; OR an Oracle (2.8 GW) or Nebius (328 MW) install timeline that slips; OR a hyperscaler publicly choosing gas turbines over Bloom SOFC.
• Theme rollover: industrial-power-AI flips SATURATED with VST/GEV/OKLO breaking down together exit signal regardless of Bloom-specific news.

Correlation Notes

Bloom trades as a high-beta member of the industrial-power-AI complex: VST, CEG, GEV, OKLO and NRG on the generation side, plus AI-data-center names IREN and Nebius, all leveraged to the same AI-capex/grid-bottleneck narrative and indirectly to NVDA build-out spend. At beta ~3.75 it moves 2-4x SPY, and in a capex unwind the basket correlation compresses toward 1 this is not a diversifier against other AI-infrastructure exposure. The idiosyncratic upside lever is SOFC-specific deal flow (repeat/expansion orders from Oracle and Nebius, or a new hyperscaler). The idiosyncratic downside lever is substitution risk a gas-turbine vendor (GE Vernova) winning a marquee data-center power mandate would hit Bloom harder than the broader power basket.