Bloom Energy's Fuel-Cell Boom Runs Into a Scandium Wall

Bloom Energy rides a 33-gigawatt US power gap and 165.5% Q2 revenue growth, but a 25x sales multiple and China-linked scandium supply draw lawsuits.

The AI trade has a power problem, and Bloom Energy has spent the past year selling the solution. Data centers are devouring electricity faster than utilities can string new transmission lines, and grid connection queues now stretch years in many US markets. That gap has turned the company's solid-oxide fuel cells — installed on-site, running on natural gas or hydrogen — into one of the more compelling infrastructure stories on Wall Street.

The market has noticed. Bloom's shares have climbed 222% since the start of the year and closed yesterday at EUR 241.50, including a 38% advance over the past 30 days alone. The rally rests on hard numbers rather than narrative: second-quarter 2026 revenue jumped 165.5% year over year to USD 1.07 billion, and earnings per share of USD 0.78 came in well ahead of analyst estimates. Management is guiding for full-year 2026 EPS of between USD 2.55 and USD 2.85.

Morgan Stanley Sees a 33-Gigawatt Hole

The demand case got a significant boost from sell-side research. Stephen Byrd of Morgan Stanley projects a 33-gigawatt electricity shortfall in the US market through 2028, and his analysis explicitly flags on-site generation as a beneficiary — with Bloom positioned as the leading supplier. The company's 800-volt direct-current fuel cells are designed precisely for power-hungry server farms, delivering immediate capacity where utilities cannot.

Institutional flows have followed. Bloom entered the S&P 500 during September's index adjustments, a promotion that reliably channels passive money into a stock. The shares also joined global benchmarks including the FTSE All-World Index and the S&P 500 Equal Weighted Index, broadening the pool of mandated buyers.

A Valuation That Assumes Perfection

None of that comes cheap. Bloom's price-to-sales ratio has pushed above 25, against a US electrical equipment industry average of barely more than twice revenue. At that multiple, the market is pricing in a scenario with virtually no room for execution missteps, project delays, or a slowdown in data-center construction.

The stock's run has been so steep that the easy money may already have been made. Index funds have done their buying, and what remains is a valuation that demands flawless delivery on expansion plans.

The Achilles' Heel in the Periodic Table

Then there is the supply chain. Bloom relies on scandium to stabilize the zirconium-oxide ceramic electrolytes inside its fuel cells — a niche input with a concentrated production base. Reports in July revealed that the company is tightly bound to Chinese supply chains through intermediaries in Thailand, Japan, and South Korea, as well as through direct deliveries.

The allegation that this dependency was concealed from the capital markets has drawn legal fire. Several US law firms have filed securities class actions, and a case is pending before the US District Court for the Northern District of California. Investors seeking to serve as lead plaintiff have until September 28 to file.

The underlying risk is not theoretical. If Beijing tightens export controls on critical metals further, the flagship project of American off-grid power could find itself stalled — a scenario that would test both the technology's promise and the stock's multiple.

A Two-Sided Bet

What emerges is a classic market dilemma. On one side sits a structural megatrend that could hardly be more real: without reliable power, the data-center boom stalls, and Bloom has the right product at the right time, proven by triple-digit growth rates. On the other, valuation metrics leave no margin for error, paired with raw-material dependencies that could halt manufacturing in a worst-case scenario.

The operational performance deserves recognition. But chasing a stock at these levels, with legal questions unresolved and a supply chain that runs through geopolitical fault lines, is a different proposition entirely.