Deep|BE: The Debate Has Shifted from AIDC Demand to Delivery Execution

Executive Summary

• AIDC power scarcity remains the structural driver. We estimate North America faces a ~15GW AI power shortfall in 2027, as grid interconnection, turbine lead times, and downstream power infrastructure constrain how much announced capacity can actually be energized. This keeps time-to-power at a premium and supports Bloom’s behind-the-meter SOFC value proposition.
• The debate has shifted from demand to execution. Production slots appear substantially booked through 2028, but the market is increasingly focused on whether Bloom can convert headline capacity into shippable MW and site deliveries. We model ~2.7GW YE27 manufacturing capacity and ~2.2GW of 2027 deliveries vs. ~2.0GW consensus, with supplier ramp, testing, yield and service capacity becoming more relevant constraints.
• Scandium looks manageable near term, but service capacity risk is overlooked. Our channel checks suggest Bloom has diversified scandium sourcing across Malaysia, Japan and Canada, while current pricing does not indicate acute scarcity. The more overlooked risk is whether Bloom’s service organization can scale quickly enough to support a rapidly expanding multi-GW installed base.
• Bloom is more than a temporary gas-turbine shortage trade. We do not expect turbine supply to meaningfully normalize by 2030, extending Bloom’s time-to-power advantage. Longer term, modularity and native-DC compatibility with emerging 800V AIDC architectures could give SOFC a structural role even as conventional generation capacity expands.