Processing, Not Mining, Is Chokepoint
Coverage of gallium and germanium shows that China dominates raw output, but the real supply bottlenecks are in specialised processing capacity rather than in mining alone. That distinction matters because it reframes supply risk into multiple operational dimensions—substrate, packaging, memory, and specialised processing—each of which should be tracked separately. Treating all supply risks as a single checkbox makes dashboards non‑actionable. (rareearthexchanges.com)
Gallium and germanium keep getting described as mining problems. That is too simple to be useful. These materials matter because they sit inside fast chips, fiber systems, infrared optics, satellite solar cells, and power electronics. But neither market really breaks at the mine. It breaks later, in the narrow industrial steps that turn awkward byproducts into wafers, substrates, epitaxial layers, tetrachloride, and packaged devices. China’s grip is real. The mistake is thinking that grip begins and ends with digging rock out of the ground. (pubs.usgs.gov) Start with the raw material itself. Gallium is mostly recovered as a byproduct of processing bauxite, and sometimes zinc ores. Germanium is also usually recovered as a byproduct, often from zinc-related streams and other base-metal processing. That means supply does not respond like copper or iron ore. You cannot simply decide to mine a lot more gallium next quarter. Output depends on what happens upstream in aluminum and zinc, then on whether someone has the equipment and chemistry to pull tiny amounts of value out of those streams. The mining headline is dramatic. The process flow is where the leverage sits. (pubs.usgs.gov) That is why China’s export controls landed so hard. Beijing’s 2023 rules did not just cover raw gallium and raw germanium. They also covered downstream forms such as gallium nitride, gallium oxide, gallium arsenide, germanium dioxide, germanium tetrachloride, and germanium epitaxial substrate material. In other words, the controls were aimed at the industrial middle of the supply chain, where substitution is slow and qualification takes time. A country can have access to ore and still be stuck if it lacks a licensed source of the right processed intermediate. (iea.org) The U.S. numbers make the point even sharper. The United States has had no domestic primary gallium recovery since 1987. It remains fully import reliant for reported gallium consumption. What it does have is a much narrower slice of the chain: one New York facility recovering and refining high-purity gallium from imported low-purity metal and new scrap. For germanium, the domestic footprint is also specialized rather than broad. A Utah company produces germanium wafers mostly for satellite solar cells from imported and recycled material. An Oklahoma company produces germanium tetrachloride for fiber-optic production, again from imported and recycled inputs. Those are not mines. They are bottleneck facilities. (pubs.usgs.gov) Once you see that, the supply map stops being one map. Gallium for GaAs wafers is not the same problem as gallium for GaN power devices. Wolfspeed, for example, emphasizes semi-insulating substrates and GaN epitaxy for RF devices up to 150 millimeters. GlobalFoundries and onsemi are pushing 200 millimeter GaN-on-silicon manufacturing for power products. Sumitomo Chemical is working on larger GaN substrates because wafer diameter itself is a constraint. These are different technical lanes with different choke points. A dashboard that marks all of them as “gallium exposed” tells you almost nothing. (wolfspeed.com) Germanium splits the same way. Fiber optics depend on ultra-pure germanium tetrachloride. Umicore says it recovers germanium from waste streams and refines it into ultrapure GeCl₄ for low-loss optical fiber. Infrared systems need optical-grade germanium with different finishing and recycling loops. Satellite solar cells need germanium wafers as substrates. These uses share a chemical family, not a common bottleneck. If fiber is fine, that does not mean infrared optics are fine. If wafer recycling improves, that does not solve tetrachloride purification. (umicore.com) Even the trade data hints at this fragmentation. In 2025, the United States imported an estimated $15 million of gallium metal and about $120 million of gallium arsenide wafers. The wafer number is the bigger clue. It shows how much value and dependency sit downstream of the metal itself. When AXT said in 2025 that longer processing times for gallium arsenide export permits were slowing substrate revenue, it was describing the real chokepoint in plain language. Not ore. Not even refined metal. Permitted, qualified, specialized material moving through a tiny number of processing steps. (pubs.usgs.gov)
