Bloom Energy, Maker of Solid Oxide Fuel Cells, Hit by Scandium Supply Concerns

Bloom Energy, the California maker of solid oxide fuel cells, is at the center of a supply chain dispute that has drawn fresh attention to Western reliance on Chinese scandium.

On July 8, 2026, activist short seller Hunterbrook published a report titled "Bloom's Big Lie" that accused the company of understating how much it depends on Chinese suppliers for scandium oxide, a material essential to its fuel cells. Hunterbrook said it used trade records, corporate filings, satellite imagery, and interviews to map four separate China-linked supply routes, and it reported that a representative of Hunan Oriental Scandium described that firm as Bloom's largest scandium supplier. The findings contradicted repeated public statements by chief executive K.R. Sridhar, who has said since early 2025 that Bloom carries no China supply chain for scandium. The report sent Bloom shares down roughly 12 percent that day. Hunterbrook disclosed that it holds a short position in the stock.

The following day, on July 9, Bloom pushed back hard. The company called the accounting and supply chain claims false and misleading, urged investors to rely on its audited financial statements and its filings with securities regulators, and said it has enough visibility across its supply chain to support annual fuel cell output of up to 25 gigawatts. Bloom also stated that it holds adequate scandium oxide inventory for current operations and planned growth and that it sources the material from several countries rather than from China alone. Shares recovered by about 6 to 7 percent after the rebuttal.

Why scandium matters here comes down to chemistry. Bloom stabilizes the ceramic electrolyte in its fuel cells with scandium, which lets the cells run efficiently at lower operating temperatures and last longer. A single 100 kilowatt Bloom server contains an estimated 13 to 15 kilograms of scandium oxide. That makes the company unusually exposed to any squeeze on scandium supply.

Scandium oxide is rarely used on its own. It is mostly a refined intermediate that gets turned into metal or blended into alloys and ceramics, and its value comes from a few high-performance uses.

The best known is aluminum-scandium alloy, where adding less than 1 percent scandium sharply improves strength, corrosion resistance, and weldability without adding weight, delivering roughly 15 to 30 percent weight savings over conventional aerospace aluminum. The F-35 fighter already uses these alloys, and they are a candidate for next-generation aircraft and hypersonic structures.

The second major use, and the largest by volume today, is solid oxide fuel cells. That is precisely the technology Bloom Energy builds, which is why the company has become the single largest scandium consumer in the world.

A third and fast-growing use is in electronics, where scandium-based thin films and substrates go into the high-power radio-frequency filters used in 5G networks.

Smaller applications include stadium and industrial lighting, catalysts, and early medical implant research.

The common thread is that scandium is a performance material with no easy commercial substitute in its most demanding roles, and that lack of a fallback option is exactly what makes heavy reliance on Chinese supply so sensitive.

The heart of Hunterbrook's argument was a supply and demand gap. Bloom aims to scale from around 1 gigawatt of installations in 2026 toward 5 gigawatts of annual capacity, a target the short seller estimated would require roughly 220 tons of scandium oxide per year. Global production sits near 240 tons, so the report argued that Bloom's growth plan is close to the physical limit of what the world produces. Bloom's response, pointing to visibility for far larger output, directly challenges that math.

The dispute lands during a genuine tightening of the scandium market. China controls an estimated 85 percent of refined scandium exports, and in April 2025 it added scandium metal, oxide, alloys, and compounds to its rare earth export controls, requiring case-by-case licenses. As of mid 2026 those controls remain in place, with Beijing issuing licenses selectively. Scandium is also awkward to source because it is almost always recovered as a byproduct of other mining rather than mined on its own, which keeps supply thin and prices volatile. High purity material has traded across a wide band, and any interruption at origin can push prices sharply higher.

Bloom's wider bet is on artificial intelligence. Demand for on-site power at AI data centers has driven a surge in orders, and the company committed roughly 100 million dollars to double annual manufacturing capacity from 1 gigawatt to 2 gigawatts at its Fremont, California, plant, with completion targeted by the end of 2026. Bloom stock has climbed sharply over the past year on that growth story, which is exactly why the durability of its scandium supply has become such a sensitive question.

Western governments have been moving to reduce the same dependence that Hunterbrook flagged. In the United States, the Defense Logistics Agency awarded Rio Tinto a 40 million dollar sole source contract for scandium oxide refined in Quebec, at a price far above historical commercial levels. The Department of Defense has also funded domestic projects, including a NioCorp facility in Nebraska targeting more than 100 tons of annual output and an ElementUS effort in Louisiana to recover scandium from industrial waste. In Europe, the ScaVanger project aims to produce scandium oxide from titanium dioxide waste. Bloom, for its part, says its multi-country sourcing and inventory buffers already insulate it from Chinese supply risk.

For now the situation remains open. Bloom is scheduled to report quarterly results on July 28, 2026, which investors will read closely for any detail on supply chain exposure and margins. The competing claims about how much scandium the company buys from China, and whether its expansion targets are realistic, have not been independently resolved.