May 16, 2026 
Carl Hirndrehe Series: Part 2 of 4: What Is Burning in There — and Why Not Already? · Part 4 of 4: Five Billion for Eternity — or 38 Billion Barrels for Forty Years
Part 2 looked at the plasma and the stellarator. Now comes the less elegant part of the story: the components.
A fusion power plant is not a gleaming reactor that only needs to be made large enough. It is a system of magnets, materials, cooling circuits, fuel systems, maintenance technology, and a long list of interfaces. Each one has to work. Many have to work at the same time.
The question is therefore no longer only whether plasma can be confined. It is how to build a plant that can operate safely, maintainably, and economically for decades.
1. HTS magnets: the game changer with a supply problem
Magnetic coils keep the plasma away from the wall. Wendelstein 7-X uses low-temperature superconductors made from niobium-titanium, cooled to roughly minus 269 degrees Celsius with liquid helium. Future compact designs increasingly consider high-temperature superconductors, or HTS, especially REBCO tapes.
REBCO is a family of rare-earth barium copper oxides. The tapes are narrow and thin, operate at less extreme temperatures, and permit very strong magnetic fields. Stronger fields can make a reactor smaller. That is why HTS is called a game changer in fusion.
The basic technology works. Commonwealth Fusion Systems tested a 20-tesla HTS magnet in 2021. Proxima Fusion plans a stellarator demonstration coil for 2027. But ARPA-E points to a serious cost gap: commercial REBCO conductors are currently in the order of 300 dollars per kiloampere-metre, while economically viable fusion plants would require a much lower price.
Quantity is also a problem. Industry does not yet produce enough HTS tape for a large fleet of reactors. A single plant would consume a substantial share of annual output, depending on its design. This is not an ordinary procurement difficulty. It is a new industrial dependency.
Europe therefore needs more than HTS research. It needs manufacturing capacity. A pilot plant at industrial scale could cost several hundred million euros up to roughly one billion by a broad estimate. That is a scenario, not a procurement figure. It still shows why “we will scale later” is not a plan.
2. Materials research: the wall that has to survive everything
The first wall of a fusion reactor receives heat, particles, and neutrons at the same time. Fusion neutrons carry roughly 14 megaelectronvolts of energy. They can damage materials at the atomic level: embrittlement, swelling, and creep are possible reasons for replacement and downtime.
Tungsten is important because of its high melting point, especially in heavily loaded areas. It is difficult to manufacture and not automatically durable. The Karlsruhe Institute of Technology works on low-activation steels such as EUROFER. Forschungszentrum Jülich develops tungsten composites. Fraunhofer institutes investigate additive manufacturing and coatings.
The real problem comes before material selection: there is no sufficiently intense, realistic source of fusion neutrons for long qualification tests. Many samples look plausible in simulations and under other particle beams. A power plant needs data on what happens after years of exposure.
That is the purpose of IFMIF-DONES in Granada. Using a particle accelerator, it is intended to create a neutron environment more relevant to fusion materials than conventional test sources. Reliable results will take time. Anyone building a plant before that evidence exists is taking a calculated risk.
That may be a sensible political choice. It should simply be called what it is. “Bold” is not always a technical category. Sometimes it is the friendly word for “not tested enough yet”.
3. The tritium breeding blanket: the largest single risk
A deuterium-tritium plasma consumes tritium. Civilian stocks are limited. The ITER overview makes clear why tritium is a bottleneck for any fusion programme.
The reactor wall must therefore do more than remove heat. The blanket is meant to absorb fusion neutrons and use lithium to breed new tritium. The eventual aim is a closed cycle, with enough surplus to cover losses, maintenance, and the startup of additional plants.
European programmes focus mainly on two concepts.
The WCLL concept uses liquid lead-lithium as the breeding material and water as coolant. Water cooling is familiar from other power plants. The electromagnetic interaction between the liquid metal and the magnetic field makes the design demanding.
The HCPB concept uses lithium ceramic pebbles and helium as coolant. Simulations suggest high breeding ratios, but the system must deal with a fixed breeding material that cannot simply be replaced during operation.
A stellarator blanket is also three-dimensional. The geometry that helps the stellarator confine plasma makes access, cooling, and even breeding uniformity more difficult.
The crucial sentence is this: nobody has experimentally demonstrated the complete tritium cycle in a fusion power plant. Components exist. Simulations exist. A power-plant demonstration does not.
ITER is intended to provide first experience through Test Blanket Modules. Under the new ITER baseline, however, deuterium-tritium operation is not planned before 2039. A commercial stellarator aiming for the late 2030s would therefore have to rely partly on simulations and preliminary tests for its blanket.
That is the largest technical qualification to smooth timetables.
4. External tritium: the clock is running
Even a working blanket does not solve the first startup. Every new reactor initially needs tritium from outside. A significant share of today’s civilian supply comes from Canadian heavy-water reactors. Quantities are small, the reactors are ageing, and tritium has a half-life of 12.3 years.
A fusion industry cannot rely on a few existing plants somehow bridging the transition. Additional production routes are needed. Researchers are examining accelerator-based concepts in which lithium is converted to tritium with neutrons. Reactor concepts designed to produce a deliberate tritium surplus have also been proposed.
Proposed is not the same as built. Depending on the design, a European tritium facility could cost from the high hundreds of millions to roughly one billion euros. Above all, it would need time for licensing, construction, and commissioning.
If a demonstrator needs tritium in 2038, 2026 is not early for a decision. It is late.
5. The divertor: one area with good news
The divertor removes helium ash and impurities from the plasma. It is among the most heavily loaded components.
The situation here is comparatively encouraging. Wendelstein 7-X has tested its island-divertor concept under conditions relevant to later power-plant operation. Water-cooled modules carried the heat loads; controlled detachment can greatly reduce the load on the surface.
That does not settle every materials question. Neutron exposure in a power plant remains. But the divertor shows that fusion research is not only a list of open problems.
6. Heat extraction and electricity generation
Steam, turbines, and generators are not unknown technologies. The challenge is integration. The blanket, divertor, and structural components deliver heat at different temperature levels. That heat must enter a stable cycle without allowing maintenance of the reactor to bring down the power plant.
Kyoto Fusioneering is developing UNITY, an integrated test facility for heat extraction, electricity generation, and tritium technology. A stellarator has a practical advantage here: continuous operation promises a steadier heat flow. A pulsed tokamak needs additional solutions for its operating cycles.
The eventual cost of a fusion plant is difficult to estimate. Studies span a very broad range. That is less a sign of secret pricing than a sign of a technology without an industrial reference plant. Anyone quoting a figure to two decimal places probably trusts their spreadsheet more than the evidence.
The total calculation
For a European “Apollo programme for fusion”, a political scenario in the order of five to ten billion euros over 15 years emerges:
- ongoing research, Wendelstein 7-X operations, and materials development: roughly three to five billion euros over 15 years;
- HTS manufacturing and industrial scale-up: roughly one to two billion euros;
- tritium infrastructure: about 0.5 to one billion euros;
- demonstration plants, partly privately financed: about one to two billion euros.
This is not a guarantee of a power plant. It is an order of magnitude for the question of whether Germany and Europe want to develop the open components at the same time.
The priorities are still clear. First, tritium infrastructure, because the lead time is long. Second, HTS manufacturing, because there is no small and affordable reactor without affordable magnets in sufficient quantity. Third, blanket and materials testing, because that is where the greatest technical uncertainty lies.
What actually slows things down
None of these problems is a fundamental contradiction in physics. Each can become a time and cost factor. Fusion may arrive in 15 years. It may also arrive in 40. The difference will not come from one spectacular breakthrough but from many medium-sized problems being addressed early and together.
That is an inconvenient message for politics and industry. It cannot be solved with another record photograph. Production capacity has to be built, test facilities financed, and decisions made before the need becomes visible in a press conference.
Part 4 therefore looks at the real bottleneck: society. Not because technology is unimportant, but because without long-term decisions technology remains a collection of very impressive prototypes.
