Isotopes & quantum materials

The materials behind this chokepoint are measured in kilograms, and that is exactly why the supplier list fits on one hand. Isotopically purified silicon-28 — the "zero-spin" substrate that gives silicon spin qubits dramatically longer coherence than natural silicon — was for two decades sourced from essentially one place: Russian centrifuge capacity originally spun up for the Avogadro kilogram project. Sanctions turned that into a Western sovereignty problem, and only two credible replacements exist. ASP Isotopes has signed three purchase agreements for enriched Si-28 — a major US semiconductor company, a global industrial-gas company, and a large US buyer — and after restarting the first 18 upgraded stages of its Si-28 enrichment cascade in May 2026 — the remaining stages get the upgrade next — has guided first commercial shipments to Q3 2026. Silex Systems is building the only other Western line: a laser-separation module at Lucas Heights, funded by an A$5.1M Defence Trailblazer grant in partnership with Silicon Quantum Computing (the initial offtake partner), guided to 5–10 kg/year from the first production module — up to 20 kg/year on Silex's product page — at the sub-500ppm Si-29 purity its current spec deems viable.

Niobium is more concentrated still. CBMM supplies roughly three-quarters of the world's niobium from a single deposit at Araxá, Brazil, and is one of only a few producers globally (alongside ATI's ex-Wah Chang operation in Oregon, Tokyo Denkai, and Ningxia OTIC) qualified for the RRR>300 high-purity grades that superconducting RF cavities and qubit films demand. In low-temperature superconducting wire, Bruker consolidated the Western merchant NbTi/Nb3Sn base in 2016 by buying Oxford Instruments' OST plant for $17.5M — the purchase price tells you how small this market is, and how little anyone else wanted the capability.

Why the moat holds. Breaking in requires either an enrichment cascade — a multi-year build with laser or aerodynamic separation IP and nonproliferation adjacency (ASPI's process heritage traces to South Africa's enrichment program; Silex's to uranium) — or metallurgy that takes a decade to qualify: high-RRR niobium and LTS wire get certified through national-lab programs (Fermilab, Jefferson Lab, ITER) on cycles measured in years. The perverse protection is the TAM itself: nobody rationally builds a cascade to chase a market measured in tens of kilograms a year. Incumbents own the choke because the prize is currently too small to fight over — which is exactly what changes if fault-tolerant roadmaps scale.

Be honest about the dollars. Quantum-attributable revenue here is tiny today. ASPI's Si-28 contracts are likely single-digit millions at a company whose shipment timing has already slipped once; Bruker's wire and ATI's niobium overwhelmingly serve MRI, NMR, fusion, and accelerators. This theme is optionality on a real physical monopoly, not a current earnings stream.

What forces the repricing. Three dated things. ASPI's first commercial Si-28 shipments, guided to Q3 2026, convert monopoly narrative into recognized revenue. DARPA's QBI Stage C down-select, expected around late 2026, pushes the 11 surviving Stage B teams toward hardware IV&V — meaning real bills of materials and multi-year offtakes with the only qualified suppliers. And the National Quantum Initiative Reauthorization Act cleared Senate Commerce unanimously in April 2026 with an adopted amendment standing up a Manufacturing USA institute for quantum manufacturing — federal scaffolding under domestic component supply broadly, even though the amendment text names no specific material.