Bloom Energy

▲ Bull Case

• ◆AI hyperscaler power crisis is structural and durable. Microsoft/Google/Amazon/Meta committed $200B+ in AI infrastructure capex 2024–2026; grid interconnection is 3–7 years; Bloom delivers in 12–18 months. Nebius deal proves contract scale is real; pipeline reportedly contains additional multi-hundred-MW opportunities.
• ◆Service + electricity recurring revenue compounds toward 35–40% of mix by FY2028, mechanically driven by installed base—every MW deployed generates ~$60K/yr for 7+ years. Combined with rising gross margin trajectory (FY2023 26.7% → FY2024 Q2 28.4%), creates credible path to infrastructure-services multiple re-rating.
• ◆Technology moat is genuine (1,400+ patents, 20-year stack development, no commercial SOFC competitor at scale) with real switching costs (proprietary replacement parts, 5–10 yr contracts, customer-specific monitoring systems)—defensible 5–10 year window during AI buildout.

▼ Bear Case

• ◆Valuation is disconnected from any defensible DCF. Step 14 Case B (using mgmt's own $3.5B FY2026 guidance) yields $10–17 per share; Case C (ultra-bull) yields $95–125. Spot of $285 is ~2x ultra-bull and 17–28x the management-guidance case. De-rating from 90x EBITDA toward industrial comps at 10–14x represents ~85% downside before business deteriorates.
• ◆Korea concentration (~25–35% of revenue) is structural binary risk. SK ecoplant balance sheet pressure in 2024 created order push-outs. A 50% Korea cut would reduce FY2027 revenue ~10–15% and damage the SK strategic narrative. Not in consensus estimates; not in spot multiple.
• ◆Dilution is structurally embedded—SBC runs ~$100M/yr (~3% of share count annually) plus 2028 converts add ~22M shares (~7% pop). The share-count-growth-offsets-EPS-growth pattern observed for 5+ years is unlikely to change before FY2028.

“Sell-side consensus PT is $237 versus spot of $285. The primary debate is NOT bull-vs.-bear on the business; it is whether the AI infrastructure premium multiple holds. Bulls argue Bloom should trade as 'Nvidia of behind-the-meter power' at 25–35x forward EBITDA; bears argue it is fundamentally a high-quality industrial equipment maker and should de-rate toward Cummins/Generac multiples (10–14x). The Nebius deal validated bull qualitative argument but quantitative multiple compression risk has not been retired. Critical secondary debate: durability of FY2026 guidance ($3.4–3.8B vs. consensus $2.5–3.2B)—a significant gap that resolves in next 2–3 quarters of order announcements.”

• ◆Additional hyperscaler contracts ($1B+) within 12 months—Microsoft, Google, Amazon reported in active discussions. Single contract announcement comparable to Nebius would validate sustained order velocity.
• ◆GAAP profitability achievement (FY2026–FY2027)—first GAAP positive year would expand investor base and shift narrative from hardware to infrastructure compounder.
• ◆Bloom Electrons reaching $200M+ annually by FY2026—validates re-rating to energy-as-a-service multiples and demonstrates recurring revenue model scalability.
• ◆Commercial-scale hydrogen electrolyzer (SOEC) orders—first 10–50 MW commercial order beyond pilots opens hydrogen infrastructure valuation narrative.
• ◆2028 convertible refinancing without dilution—debt refinancing on favorable terms removes the convertible overhang and validates capital structure stability.

• ◆Multiple compression / de-rating—Current 90x FY2026E EBITDA multiple is unsustainable on any historical base rate for industrial-equipment hybrid companies. Even gradual de-rating toward 30x produces 60%+ stock decline before business deteriorates.
• ◆Korea/SK relationship deterioration—25–35% revenue concentration; SK ecoplant balance sheet pressure persists; concentration risk not embedded in consensus.
• ◆AI hyperscaler capex pause—Microsoft/Google/Amazon could defer capex if AI demand slows or capex efficiency improves; would defer Bloom order conversion.
• ◆IRA Section 48 ITC rollback—Republican Congress legislative risk; would compress customer ROI and require Bloom to absorb 15%+ price reductions.
• ◆Convertible note dilution—2028 converts are deep in-the-money (~$285 stock vs. $28.50 conversion price); base case assumes ~22M share dilution (~7%). SBC compounds 2–3%/yr; cumulative dilution material.
• ◆Technology disruption—Long-duration battery storage cost decline, SMR nuclear timelines, competing PEM advances could displace fuel cells in AI use case over 10-year horizon.
• ◆CEO succession risk—K.R. Sridhar concentrated in technology, relationships, and governance; no clear successor identified; creates execution risk if transition needed.