Bloom Energy’s Pitch to AI Data Centers: Cost Savings and Future Prospects
On September 16, Bloom Energy announced significant potential savings for AI data centers that utilize its solid oxide fuel cells. According to their cost model, a 1-gigawatt data center using Bloom’s technology could spend $3.6 billion less (27% savings) on non-chip construction compared to traditional AC power solutions. Over a five-year period, the total ownership costs could decrease by $5.5 billion, translating to a 9% reduction. This claim is substantial, especially when compared to Bloom’s projected revenue of $3.9 billion to $4.2 billion for the entirety of 2026.
Credibility and Market Context
While analysts like me find these claims credible, the real impact on Bloom’s stock valuation remains uncertain. The stock is trading near $289, approximately 18% below its peak of $351.28 within the past year. This price point emphasizes the stock’s valuation in the context of Bloom’s projected earnings and future market growth.
Breakdown of Savings
The $3.6 billion in savings primarily results from Bloom’s fuel cells generating 800-volt direct current (DC), which is the required input for AI chips. Traditional setups often necessitate several conversions from AC to DC, which can be both time-consuming and costly due to the need for equipment like transformers and switchgear. Bloom estimates that the total cost for non-chip construction on a 1-gigawatt site is about $13 billion, while total ownership costs approximate $60 billion over five years.
Competition and Market Dynamics
Nvidia plans to deploy its 800-volt DC data centers using a different approach—converting AC to DC once near the data center’s edge. This methodology might decrease the savings Bloom touts, especially if Nvidia and its partners (like Eaton and GE Vernova) establish a foothold in this niche without relying on Bloom’s technology.
Speed and Demand for Fuel Cells
Bloom’s recent success with Oracle, which expanded its partnership to include up to 2.8 gigawatts of Bloom’s fuel cells, suggests that factors like speed of deployment are crucial for securing contracts rather than just cost. Oracle’s spokesperson emphasized rapid deployment means efficiency and reliability are top concerns, which could influence future orders.
Financial Metrics and Future Consideration
In the second quarter of 2026, Bloom’s product revenue surged to $935 million, more than tripling from the previous year. However, gross margins have remained consistent, indicating that while sales volume has increased, the pricing power has not. Furthermore, Bloom needs a stable gas supply for its fuel cells; any disruptions could jeopardize ongoing projects.
The current stock valuation stands at around 56 times expected 2027 earnings. With the timeline aligned with when Nvidia’s systems are expected to roll out, the competitive landscape may shift if potential clients don’t commit to Bloom’s technology.
Conclusion
Bloom Energy’s cost savings claims for AI data centers are compelling but warrant careful scrutiny before making investment decisions. Confirmed commitments from major clients, along with a clear competitive advantage against established players like Nvidia, will be crucial to judgment on Bloom’s future stock performance.
