Tungsten melts at 3,422°C. No other metal on the periodic table except certain forms of carbon can match it [1]. That property makes tungsten the leading candidate for fusion reactor first walls, the innermost barrier that surrounds the plasma and absorbs neutron bombardment, heat, and magnetic stress [1]. It also makes tungsten nearly impossible to manufacture at scale, because most industrial metalworking processes top out below 2,000°C.
The U.S. Department of Energy's Advanced Research Projects Agency selected 13 projects in 2024 under its CHADWICK program to solve the materials challenge [1]. Ames National Laboratory leads one project and collaborates on a second with Pacific Northwest National Laboratory [1]. Both focus on the first layer of the first wall, the surface in direct contact with the fusion reaction [1].
That first layer has two requirements that do not easily coexist. It must resist cracking and erosion over years of operation [1]. It also cannot remain radioactive long enough to prevent maintenance crews from servicing the reactor [1]. Tungsten passes the first test. Whether it passes the second depends on neutron activation rates that vary with the isotopic composition of the material.
The manufacturing constraint is binding. Ames Lab acquired a commercial platform for producing refractory materials, metals that resist deformation at extreme temperatures, in late 2024 [1]. By spring 2025, the lab added two more systems, one for lab-scale quantities and one for pilot-scale production [1]. Nicolas Argibay, the Ames scientist leading one CHADWICK project, confirmed the timeline [1]. The fact that a national laboratory is building this capability from scratch in 2025 reveals how far fusion materials are from commercial readiness.
The first wall is not a single material. It has two layers with incompatible jobs [1]. The inner layer, where tungsten is concentrated, must not crack or stay radioactive [1]. The outer layer must extract heat for electricity conversion while withstanding magnetic fields strong enough to contain plasma [1]. The interface between these layers, where a refractory metal meets a heat-transfer alloy, is its own engineering problem. Thermal expansion rates differ. Bonding methods that work at room temperature fail at 1,500°C.
Ames Lab has also invested in equipment to measure mechanical properties of refractory materials at what the lab calls "relevant temperatures" [1]. Relevant means near the melting point of tungsten. The fact that this measurement infrastructure is new in 2025 shows the gap between laboratory samples and reactor-scale components. A fusion plant will require tons of tungsten in precise geometries, tubes, tiles, layered composites, all of which must perform identically under neutron flux. No one has made that much tungsten in those forms.
The duplication across CHADWICK projects signals uncertainty. Thirteen teams are working on first-wall materials [1]. Ames leads one tungsten effort and collaborates on another [1]. If the path forward were clear, the program would not fund parallel approaches to the same layer of the same component.
Two systems by spring 2025 [1]. Pilot-scale.
Two systems by spring 2025 [1]. Pilot-scale. That is not mass production; it is the stage before someone knows whether mass production is possible. The timeline from pilot equipment to a functioning supply chain is measured in decades, not years, and fusion needs that supply chain before the first commercial reactor breaks ground.
